Archive for AI

AI Mimics CEO’s Voice To Steal Over £200,000

A recent Wall Street Journal report has highlighted how, in March this year, a group of hackers were able to use AI software to mimic an energy company CEO’s voice in order to steal £201,000.

What Happened?

Reports indicate that the CEO of an unnamed UK-based energy company received a phone call from someone that he believed to be the German chief executive of the parent company.  The person on the end of the phone ordered the CEO of the UK-based energy company to immediately transfer €220,000 (£201,000) into the bank account of a Hungarian supplier.

The voice was reported to have been so accurate in its sound, that the CEO of the energy company even recognised what he thought was the subtleties of the German accent of his boss, and even “melody” of the accent.

The call was so convincing that the energy company made the transfer of funds as requested.

Fraudster Using AI Software

The caller, who was later discovered to have been a fraudster using AI-base voice-altering software to simulate the voice of the German boss, called 3 times.  In the first call, the fraudster requested the transfer, in the second call they (falsely) claimed that the transfer had been reimbursed, and in the third call the fraudster requested an additional payment. It was this third call that aroused suspicion, partly based on the fact that the telephone number appeared to indicate that the caller was in Austria and not Hungary.

Money To Hungary, Mexico and Beyond

Unfortunately, the money had already been transferred to a Hungarian account after the first call, and it has since been discovered that money was immediately transferred from the alleged supplier’s Hungarian bank account to an account in Mexico, and then further disbursed to accounts in other locations, thereby making it very difficult for authorities to follow the trail.

What Sort of Software?

The kind of software used in this attack may have been similar in its output to that demonstrated by researchers from Dessa, an AI company based in Toronto.  Dessa has produced a video of how this kind of software has been able to produce a relatively accurate simulation of the voice of popular podcaster and comedian Joe Rogan – see:

What Does This Mean For Your Business?

It is known that cybercriminals, deterred by improved and more robust enterprise security practices have decided to look for human error and concentrate more on social engineering attacks, a category that this voice simulation attack (via phone calls) fits into. The fact that this attack has taken place and been successful shows that some cybercriminals are already equipped with the computing power and most up-to-date machine-learning AI technology that they are clearly capable of using.

This means that companies and organisations (particularly larger ones), may now be at risk of facing more sophisticated deception and phishing attacks. The AI company Dessa has suggested that organisations and even individuals could expect to face future threats such as  spam callers impersonating relatives or spouses to obtain personal information, impersonations intended to bully or harass, persons trying to gain entrance to high security clearance areas by impersonating a government officials, and even an ‘audio deepfake’ of a politician being used to manipulate election results or cause a social uprising.

Companies should try to guard against social engineering attacks by educating all staff to the risks and having clear verification procedures (and not just relying on phone calls), tests, and chain of command authorisation in place for any requests for funds.

Record Levels of Investment in UK AI

A Tech Nation Report has shown that AI investment in the UK reached record levels in the first six months of the year making it the third biggest market in the world for AI investment, just behind the US and China.


Crunchbase figures show that AI investment in the UK reached £859.29m in just the first six months of this year, compared to £825.85m for the whole of last year.

This latest surge in AI investment marks five years of consecutive growth and a massive six-fold increase between 2014 and 2018.

Also, AI start-ups in the UK raised almost double the amount of those in the rest of Europe combined.

Why The High Investment Levels?

The AI investment growth can be attributed to several factors, not least:

  • A rise in the number of start-ups with 50 or fewer employees.  These account for 89% of the UK’s AI companies.
  • The Department for Digital, Culture, Media and Sport (DCMS) £1bn AI sector deal to put the UK at the forefront of the AI industry, including almost £300m of new private sector investment, as part of the UK government’s Industrial Strategy (announced November 2017).  This initiative was intended to establish partnerships between government and industry in order to increase productivity.


Even though the figures show that the investment trend is going in the right direction, UK-based companies hoping to make the most of AI face some clear challenges including:

  • A tech skills shortage and a so-called “brain drain” in the UK and across Europe as top university tech students are tempted to work further afield e.g. in the U.S.  Also, Brexit fears in the UK have deterred some European specialist tech workers from staying.
  • Challenges in scaling up their businesses so that they can become competitive in the global market.

Small Pool

These challenges to the growth of AI companies mean that there is only a relatively small pool of UK AI-focused companies that have been able to make the step to scaling-up and competing on the world stage.  AI companies in other countries such as China, by contrast, tend to have larger workforces e.g. 53% have more than 50 employees.

There is also a relatively small pool of people in the world who can contribute to cutting-edge AI research.

Benefits and Threats of AI

AI offers many benefits to businesses such as cost and time savings (greater productivity and reduction in errors), the ability to make better use of resources (AI handles repetitive jobs and bots handle common questions).

Many people are, however, concerned that the growth in AI will mean a loss of jobs e.g. Gartner figures show that AI could eliminate 1.8 million jobs.  It should also be remembered that AI could create 2.3 million jobs by 2020 (Gartner) and that if the large-scale introduction of AI follows the pattern of temporary job losses followed by recovery and business transformation, the combination of human and artificial intelligence could provide exciting news competitive advantages for businesses.

What Does This Mean For Your Business?

The investment in AI within the UK is promising for the tech sector, the economy, and for the future of the UK in the global tech market, provided that UK-based AI companies can tackle the challenges of being able to scale-up and successfully find the human tech talent at a time of skills shortages.

AI may cost jobs in the shorter term, but it may also bring new strengths and opportunities to businesses and could transform the way we are able to work for the better.

BBC to Launch Own ‘Beeb’ Digital Voice Assistant Next Year

The BBC has announced that it will be launching its own digital voice assistant ‘Beeb’ next year to work on all smart speakers, TVs, and mobile devices.


The new digital (AI) voice assistant, which is being developed by an in-house team, will be trained to have a good understanding of the many different UK regional accents.  This has meant that BBC staff from around the UK have been invited to record their voices to help train the programme.


Even though the BBC has not said that ‘Beeb’ will be sold with its own hardware device (smart speaker), as an AI digital voice assistant it will essentially be in broad competition with Amazon, Google and Apple, all of which have already been in the market for some time with their own voice assistants.

That said, in addition to not being released in a bundle with a home smart speakers to compete on the shelves with other general smart speakers, Beeb is different because it has been designed, rather like the iPlayer, as a means to provide easier access to the BBC’s own content, programmes and services.  It is thought that ‘Beeb’, being a BBC product that’s specifically designed with the purpose of accessing BBC content, will mean that it is trusted and used by BBC customers.

Voice-Activated Future

As a public services broadcaster, the BBC sees ‘Beeb’ as an important step to keep up with the times in what it describes as a “voice-enabled future”.  For example, 20% of British households already use voice assistants (Guardian).


Some critics have pointed out that having a single syllable word such as ‘Beeb’ as the wake-word could lead to mistakes being made by the assistant, but the BBC says that ‘Beeb’ is still just a working title.

No More BBC on TuneIn

From the end of September, the BBC’s radio stations will no longer be available through the TuneIn radio app (as used by Alexa) because it has been reported that Amazon will not share information about listeners of BBC stations.

What Does This Mean For Your Business?

The new digital voice assistant is a way in which the BBC can give its customers a more convenient and modern way to access its content, in the same way that some competitors are using  Netflix uses Amazon voice controls on Fire TVs, and at a time when people are used to using other voice assistants. Also, Beeb is a part of the BBC’s move to push users towards its own products, and crucially, to find out more information about its users.  This has been shown, for example, by the need to sign-up to view programmes on iPlayer, and by the impending removal of BBC stations from TuneIn app over a lack of information-sharing.  The BBC’s own digital assistant will mean that it can have information-gathering systems built-in.  This, in turn, helps the BBC to better target its services and to compete more effectively in the wider marketplace, while at the same time, help it to improve and add value to its public service broadcasting.

Four-Year Lifespan For Self-Driving Cars

As large car manufacturers seek to reinvent themselves as ‘mobility companies’ in an effort to compete for global leadership in the growing autonomous driving sector, a Ford Executive has predicted that self-driving cars will only last four years.

Only Four Years?

The prediction of four-year lifespan for self-driving cars came from John Rich, the operations chief of Ford Autonomous Vehicles, in a recent interview with the Telegraph.

Why Four Years?

The idea that a driverless car will only last four years stems from the fact that these cars will be part of fleets that have continuous use and will, therefore, wear out more quickly.  Even though this may appear to indicate that car companies could make more money by selling new car replacements after only four years, this is not necessarily so because car manufacturers appear to envisage a future where they will become fleet operators that sell us fewer cars.

Mobility Company

Mr Rich’s prediction fits in with the idea that traditional car manufacturers such as Ford and Toyota say that they’re aiming to become ‘mobility companies’ that operate fleets of autonomous/driverless vehicles for other companies to use.  This could include the car manufacturers hiring the fleets out themselves, supplying the fleets for other companies to hire out, and getting involved in ventures with other operators.  For example, Toyota and Chinese autonomous driving company have recently teamed up in a US$600 million joint venture to explore mobility services and to help Toyota to become a major mobility company in China. Also, Pittsburgh start-up Argo AI is reported to be developing driverless cars for Ford and is testing the technology in five cities in the US.

The move by Ford and other manufacturers towards becoming mobility companies with autonomous fleets will see them compete directly with operators such as Uber.

Decline In Private Ownership

The prediction and vision from market analysts is that there will be a decline in private car ownership and the costs associated with that as consumers will prefer to use the widely available fleets of autonomous vehicles operated by the new mobility companies.

What Does This Mean For Your Business?

Traditional car manufacturers appear to see their future as mobility companies in a world where they and other businesses operate fleet services of widely available autonomous vehicles to business and individual users who will no longer need to own a car themselves. This is all part of today’s car manufacturers trying to get significant peace of global (in the developed world) market for autonomous transport.  If this future vision plays out as the car manufacturers and analysts predict, this will have a dramatic effect on businesses and markets along the car supply chain as well as the private hire and public transport markets.

Grammar Correction Capabilities For Gmail and G Suite

Google has announced that it is rolling out new real-time, AI-powered spelling and grammar correction capabilities for G Suite users and personal Gmail accounts.

Real-Time, As-You-Type

The real-time, as-you-type spelling autocorrection is now available to all G Suite users and personal Gmail accounts, but the Grammar suggestion feature is only available to G Suite users.

Autocorrection Spelling

The new features use AI to spot and highlight mistakes with spelling and grammar, and even the use of tenses.

The autocorrection spelling feature uses coloured (red) squiggly underlines to highlight spelling mistakes, the mistakes can be auto-corrected, and the changes are temporarily underlined in grey so that you can see the difference between the two.


The AI-powered grammar correction feature, which has previously been available in Google Docs, uses a squiggly blue line to highlight errors in grammar, and pop-ups appear to help make changes to the grammar and allow the user to mark them as correct or switch them back to how they were.


The autocorrect spell-check feature has been added as a default and if users want to turn it off, they can do so by going to Gmail’s Settings > General and then turn off Grammar, Spelling, and Autocorrect.


The benefits to users could be the ability to work smarter and faster, make a better impression with their communications, and learn more about correct spelling and grammar and use the knowledge to feel more confident in their writing over time.


These new features are, however, in competition with, and could affect profits for other third-party spelling and grammar services e.g. Grammarly which relies upon users opting to use an upgrade to a freemium service.

Texting To Blame?

Ofcom figures showing that teenagers, young adults and adults chose text-based communications as their preferred way of communication highlight one of the reasons why many people believe that mobile phone usage, with its reliance on auto-correct, is partly to blame for a decline in the standards of spelling and grammar that has necessitated a further need for more sophisticated AI-based tools.

What Does This Mean For Your Business?

For third-party spelling and grammar apps such as Grammarly, this move by Google could reduce the value of (and the perceived need for) their services.

For business users of Google’s services (G Suite and Gmail) these kinds of tools could help save time and improve the quality and consistency of their communications which in turn could positively reflect on their brands.

For Google, these new features could provide an improved experience for their users and add more loyalty and perceived value to Google’s range of services.

Some critics have, however, noted that these features could homogenise the way that we write, could raise privacy concerns, and that the AI technology could also be subject to bias in its suggestions.

Opting Out of People Reviewing Your Alexa Recordings

Amazon has now added an opt-out option for manual review of voice recordings and their associated transcripts taken through Amazon’s Alexa but it has not stopped the practice of taking voice recordings to help develop new Alexa features.

Opt-Out Toggle

The opt-out toggle can be found in the ‘Manage How Your Data Improves Alexa’ section of your privacy settings, which you will have to sign-in to Amazon to be able to see.  This section contains a “Help Improve Amazon Services and Develop New Features” section with a toggle switch to the right-hand side of it and moving the toggle from the default ‘yes’ to the ‘no’ position will stop humans reviewing your voice recordings.

Echo owners can see the transcript and hear what Alexa has recorded of their voices by visiting the ‘Review Voice History’ of the privacy section.

Why Take Recordings?

Amazon argues that training its Alexa digital voice assistant using recordings from a diverse range of customers can help to ensure that Alexa works well for all users, and those voice recordings may be used to help develop new features.

Why Manually Review?

Amazon says that manually reviewing recordings and transcripts is another method that the company uses to help improve their services, and that only “an extremely small fraction” of the voice recordings taken are manually reviewed.

Google and Apple Have Stopped

Google has recently been forced to stop the practice of manually reviewing its auto snippets (in Europe) by the Hamburg data protection authority, which threatened to use Article 66 powers of the General Data Protection Regulation (GDPR) to stop Google from doing so.  This followed a leak of more than 1,000 recordings to the Belgian news site VRT by a contractor working as a Dutch language reviewer.  It has been reported that VRT was even able to identify some of the people in the recorded clips.

Apple has also stopped the practice of manual, human reviewing of recordings and transcripts taken via Siri after a (Guardian) report revealed that contractors used by Apple had heard private medical information and even recordings of people having sex in the clips.  This was thought to be the result of the digital assistant mistaking another word for its wake word.

What Does This Mean For Your Business?

If you have an Amazon Echo and you visit the ‘Review Voice History’ section of your privacy page, you may be very surprised to see just how many recordings have been taken, and the dates, times, and what has been said could even be a source of problems to those who have been recorded.  Even though we understand that AI/Machine Learning technology needs training in order to improve its recognition of and response to our questions, the fact that mistakes with wake words could lead to sensitive discussions being recorded and listened to by third-party contractors, and that voices could even be identified from those recordings highlights a real threat to privacy and security, and a trade-off that many users may not be prepared to accept.

It’s a shame that mistakes and legal threats were the catalysts for stopping Google and Apple from using manual reviewing, and it is surprising that in the light of their cases, Amazon is not stopping the practice as a default altogether but is merely including an opt-out toggle switch deep within the Privacy section of its platform.

This story is a reminder that although smart speakers and the AI behind them bring many benefits, attention needs to be paid, as it does by all companies to privacy and security when dealing with what can be very personal data.

A.I. Powered Bar-staff. Who’s Next?

In what’s been called the world’s first ‘A.I. Bar’ (developed by British data science product company DataSparQ) ordering a drink at a busy bar has been made easier, faster and fairer by using facial recognition technology to place customers in an “intelligently virtual” queue.

Solving Old Problems

Information and statistics (DataSparQ) show that pub-goers in Britain spend more than two months over a lifetime queuing for drinks and that people pushing in at bar queues is the biggest gripe.  Who to serve next as efficiently as possible without causing an argument, and how to spot underage customers at busy times are challenges faced by many bar workers.  Also, solo drinkers and females can find busy bars intimidating and frustrating.

The new DataSparQ ‘A.I. Bar’ Software-as-a-Service product, which costs landlords from just £199 a month and uses a standard webcam, display screen and Internet connection to link up to A.I facial recognition technology appears to be able to address all of these challenges.

How It Works

The A.I. Bar, which has been tested in London, uses a camera linked to the machine learning technology to spot those persons arriving at the bar.  The system displays a live video of everyone queuing on a screen above the bar and a number, which appears above each customer’s head, representing their place in the queue. The system also protects customer privacy by deleting the data (pictures of faces) within 24 hours.

For bar staff, the ordered numbering of customers, and the fact that customers are clearly aware of their number in the queue reduces the chance of arguments. The system shows the bar staff on an iPad, exactly who to serve next thereby helping bars and pubs to maximise their ordering efficiency. The system also tells bar staff who they should ask for I.D. to verify their age, thus helping the pub/bar to stay on the right side of the law.

More Pints Served

In tests of the system, the before and after data has revealed there was an overall reduction in serving times with equivalent of more than 1,600 pints extra poured over a year compared to the average UK pub.  This could equate to a potential 78million additional pints poured a year if the UK’s 48 thousand pubs adopted the A.I Bar technology.

What Does This Mean For Your Business?

For UK pubs and any business which have to deal with busy bars (hotels, clubs, live music venues and festivals), this system is an example of how the latest technology can be used in a practical setting to solve a number of age-old problems that have troubled drinkers, owners and staff alike.  If this system was widely adopted, the efficiencies created, the extra beer sales, and the reduction of potentially intimidating situations in pubs could benefit the wider pub and drinks trades, and could go some way to helping at a time when so many pubs are being forced to close.

UK National Surveillance Camera Day – 20th June

In a world first, the UK is playing host to an awareness-raising National Surveillance Camera Day 20 June as part of the National Surveillance Camera Strategy.

National Surveillance Camera Day

The National Surveillance Camera Day, which is part of the UK government’s National Surveillance Camera Strategy for England and Wales consists of events around the country that are designed to raise awareness, inform and lead to a debate about the many different aspects of CCTV camera use (and facial recognition use) in the UK. The Surveillance Camera Commissioner (SCC) is hoping that the public will take the day as an opportunity to have their say about the future of surveillance cameras with the regulators and service providers listening.

It is hoped that points raised in the debate triggered by the day could help inform policymakers and service providers about how the public feels about surveillance practices and how surveillance camera system use fits with society’s needs and expectations.

One of the key events to mark the day is the “doors open” initiative to allow the public to see first-hand how surveillance camera control centres are operated at the premises of signatories to the initiative e.g. local authorities, police forces, hospitals, and universities.

What / Who Is The SCC?

The Surveillance Camera Commissioner (SCC) for England and Wales is appointed by the Home Secretary as set out in the Protection of Freedoms Act 2012 (PoFA) and it is the Commissioner’s role to ensure surveillance camera systems in public places keep people safe and protect and support them. The current SCC is Tony Porter.

What Is The National Surveillance Camera Strategy?

The National Surveillance Camera Strategy is the government document, presented by the SCC that outlines the plans for surveillance camera use going forward.  The 27-page document is available online here:

Two Related World Firsts

Another related world first that is due to take place on the same day as National Surveillance Camera Day will be the launch by the SCC of a “secure by default” list of minimum requirements for manufacturers of video surveillance systems, designed for manufacturers by manufacturers.  The hope is that where manufacturers meet the new “secure by default” minimum requirements, this should ensure that the default settings of a product are as secure as possible, and therefore less likely to be vulnerable to cyber-attacks that could lead to data breaches.

What Does This Mean For Your Business?

Most of us are used to (and often no longer notice) CCTV cameras in use in business premises and public spaces, and we accept that they have a value in protecting us and our businesses in terms of deterring criminals and playing an important role in identifying them, and in providing valuable evidence of crime.

Holding a National Surveillance Camera day highlights the fact that new and emerging technologies e.g. facial recognition and AI are currently causing concern in terms of possible infringements to civil liberties, privacy and security, and an ‘open-day’ style approach could have benefits both ways.  For example, it could serve to reassure the public and at least let them feel that their views and concerns will be listened to, while at the same time giving policy-makers an opportunity to gauge public opinion and gather information that could help guide their strategy and communications.

It is good news that manufacturers are setting themselves minimum security standards for their CCTV systems as part of “secure by default”, as this could have knock-on positive effects in protecting our personal data.

Revenue Risk To UK Companies Too Slow To Adopt AI

Research from the McKinsey Global Institute shows that UK companies could lose 20% of their cash flow if they are too slow to invest in and adopt Artificial Intelligence (AI) tools.

Could Miss Out

Even though the report highlighted the UK’s higher than average AI-readiness, the country could miss out on a potential 22% boost to the economy and a 120% growth for individual businesses if organisations do not start investing now in AI tools that could help them gain considerable competitive advantages.

Investment Pockets

The research noted that the UK currently only has pockets of innovation for AI e.g. Google’s DeepMind AI division, and that in order to replicate this kind of innovation for growth, businesses need to be in a position where they can offer AI at scale, invest in the necessary talent and find ways to use the findings of the latest research to help achieve commercial success.

IT Skills Shortage

The UK already has an IT skills shortage and is experiencing a “brain drain” from UK university talent to US companies, a further brain drain pressure caused by Brexit fears, and the pull of attractive higher salaries and advanced tech sector careers in tech firms overseas.

Oxford University – Massive Donation For AI

One way of combatting an AI brain-drain and helping to grow UK AI talent which could help UK businesses with AI is to have an AI centre in the UK.  Oxford University has just received the largest single donation to a UK university of £150m from US private equity billionaire and Republican political adviser Stephen Schwarzman for the purpose of building an institute to study the ethics of AI. Mr Schwarzman is reported as saying that artificial intelligence is the major issue of our age.  He has also given £279m to the Massachusetts Institute of Technology (MIT) to establish a centre for computing and artificial intelligence.

What Does This Mean For Your Business?

Both this research, and some Nesta research from last year have highlighted how UK businesses may be facing added competitive challenges and missing out on revenue in the not-too-distant future due to an ongoing skills shortage that has been amplified and exacerbated by Brexit uncertainty, and by late investment in and adoption of AI.

McKinsey’s latest research builds on its research from last year where it attempted to simulate the effects of AI on the global economy.  The results showed that AI could deliver additional global economic activity of around $13 trillion by 2030, or about 16 per cent higher cumulative GDP compared with today. This, of course, would be good for businesses that have invested in AI, and where many of the potential challenges are adequately tackled e.g. the UK’s IT skills shortage.  It should also be accepted that the productivity growth that AI could help fuel is likely to be affected by a host of different factors in different parts of the world e.g. labour automation, innovation, the pace of adoption of AI, and the global connectedness or labour-market structure of any given country.

It is also worth noting that AI can deliver threats as well as opportunities, in the form of AI-based cyber-attacks which are a developing risk to whole nations and economies as well as individual businesses. This is certainly one area where nations such as the UK must invest in its own AI defence structures and tools.

Mastercard’s AI-Based Digital Wellness Could Make Online Purchasing Easier and Safer

Mastercard has announced the introduction of its Digital Wellness program which utilises AI-based click-to-pay technology and new standards in order to provide an easier and safer online shopping experience.

The Program

The Mastercard Digital Wellness program provides tips and resources that are designed to help businesses (especially small and independent businesses) protect themselves from cyber-attacks and data breaches. The program includes Secure Remote Commerce, Mastercard’s Cyber Readiness Institute (a collective of business leaders), and The Global Cyber Alliance which provides SMBs with free cyber-security tools.

New Click-To-Pay Checkout System

Coming out of the Digital Wellness Program is Mastercard’s new click-to-pay checkout system which is enabled by Mastercard’s deployment of EMVCo’s (Europay, Mastercard, Visa) specification. The standards that make up EMVCO’s specification provide a foundation that enables the processing of e-commerce transactions in a consistent, streamlined fashion over a variety of digital channels and devices, including smartphones, tablets, PCs and other connected devices.

This means that the click-to-pay checkout system can be used for all kinds of online shopping, across multiple devices, and across cards, and can replace old key-entry checkout systems.

Tokenization and NuData

The click-to-pay checkout system incorporates tokenization and NuData, which represent Mastercard’s AI and machine learning tech. NuData can prevent fraud by (for example) monitoring website traffic changes, analysing changes in browsers and web surfing speeds, and verifying all the user data that makes a user unique (such as an individual’s scroll speed on their device).

The inclusion of AI technology means greater security and no need for customers to enter passwords when they pay.

The Advantages

The key advantages of the click-to-pay checkout system from the Digital Wellness Program are that:

  • It tackles the problem that customers feel unease when it comes to paying for things online because of the added security.
  • It’s fast and easy – the instant click-to-pay with no need for passwords tackles the reluctance of online shoppers to create a new user account.
  • Merchants who adopt the system have a system from a known and trusted provider that could give them a better chance of preventing fraud.

These factors mean that the system could make customers more likely to feel comfortable shopping for things on smaller websites or with unknown retailers.

What Does This Mean For Your Business?

For Mastercard, this is a way of selling its services to the huge market of smaller and independent businesses.

For merchants, it’s a way for them to leverage the latest AI tech to protect themselves and their customers from fraud, and tackle popular known barriers to purchases from smaller retailers online i.e. worries about security and the unwillingness to take the time to set up a new user account when they want to buy something.

For customers, the system should provide a safe and fast purchasing experience which can only reflect well on the merchant.  It remains to be seen, however, how many merchants take up the new system and what the cost versus benefit implications will be.