Sunday, June 18, 2023

Portugal helps reduce AI risks



Artificial intelligence (AI) has become a global revolution and has brought us well into the dawn of a new era. It is at present hugely helpful in many respects, but very difficult to comprehend while almost certainly set to impose heavier and unpredictable risks as it becomes more advanced.

The deep complexities of AI will be discussed at the 1st International Conference on Explainable Artificial Intelligence, which will be held in Lisbon next month.

Much confusion and many myths surround AI, which is essentially a man-made technological machine. It got properly underway in the late 1950s and, unknown to most of us, it was developed so that by the mid-70s it enabled computers to store much more information.  AI machines were beginning to think in much the same way as their human creators, though without any emotion. All sorts of knowledge, fast problem-solving and the interpretation of spoken language became readily available to all with a mobile phone or an ordinary computer.

AI is now commonplace, a normal part of our everyday lives. It is central to our casual or in-depth searches for information. It is good at quickly finding solutions for problems that non-AI machines are incapable of coping with. Office workers access more information more quickly and thus sometimes reduce working hours or staff numbers. Companies also profit from AI’s ability to allow us to arrange travel plans, shop online, check the best medications for specific purposes, or simply apply for job offers or citizenship. Such advanced conceptions as safe AI cars with no need for a driver at the wheel are well on the way.

However, there are growing concerns about possible pragmatic and ethical matters. Leading experts have different views on how these risks may further develop and what can be done to control them. This emphasises the importance of next month’s conference in Lisbon (July 26-28), as well as a number of other AI conferences and workshops in Portugal, both this year and next. They will bring together various academic specialists from Europe, North and South America, Asia, Africa and Australia. Hopefully, these and similar get-togethers across the world will help wise up the rest of us on the precautions we need to take now and in the future to better understand and update the changes of AI.

During Rishi Sunak’s recent visit to Washington he announced that the UK would host the first global summit on AI regulation later this year. At the recent G7 summit in Japan the group discussed and set in motion the creation of an intergovernmental forum called the “Hiroshima Process”.  Other major recent events have included an open letter with hundreds of signatures from some of the biggest names in technology and AI’s most distinguished academics.

In June 2019, the Portuguese government presented an ‘AI Portugal 2030’ strategy aimed at fostering the best public and private use of artificial intelligence throughout the next decade. Under Portugal’s presidency of the European Union in 2021, the focus was on the adoption of the first EU law to maintain transparency and respect for AI users’ rights.

This week, the European Parliament produced a landmark draft law that is the world’s most far-reaching attempt to address the potentially harmful effects of artificial intelligence. If passed, the law will be an example to leaders around the world of how to extensively control AI, or in the words of the New York Times, “to put guardrails on the rapidly expanding technology.”

Among other things, the EU’s draft AI Act would ban or severely restrict the use of technology in biometric surveillance such as facial recognition software, while requiring makers of AI systems such as ChatGPT to disclose all their AI-generated data content.

The draft EU law is not expected to be passed until later this year. Meanwhile, the risks already presenting themselves involve consumer data privacy, disinformation, and biased programming, as well as inadequate legal regulations. Many companies simply disregard data privacy as there are insufficient national and international regulations on this. AI is only as correct and unbiased as the data fed into it by human programmers.   

Future issues could include AI robots being programmed to cause different sorts of serious damage. The technologies could be used to provide benefits to one side or the other in major international disputes or warfare.

Much of the control of AI will be down to the expertise and management of the younger generation, which is why the AI Portugal 2030 initiative is so important. It seeks to support how students and young graduates grapple with the complexities of this subject.

As to the notion that AI robots may eventually take over the world, thia is far from science fiction.  Many scientists, including the renowned evolutionary biologist Richard Dawkins, believe it to be very possible.

 


No comments: