Bridging the regulatory gap plus other AI stories to read this month

As the European Union pushes ahead with its AI Act, it is set to work more closely with the US government to establish minimum standards before legislation comes into force.

EU Commissioner Margrethe Vestager pointed out that while the policy process might be completed by the end of 2023, it would still take one or two years for the act to come into effect. Interim measures would be needed to bridge that gap, Reuters reports.

The Biden administration had set out its vision for AI governance in its Blueprint for an AI Bill of Rights, published last autumn, while regulatory efforts are also being pursued at state level.

EU Commissioner Vestager told Reuters that bridging the regulatory vacuum would be one of the topics at the EU-US Trade and Technology Council meeting of ministers in Sweden later this month.

This follows a call to action from the leaders of the G7, meeting in Japan on 20 May, for technical standards to keep AI trustworthy and aligned to the group's shared values. G7 ministers have been tasked to set up a working group on AI by the end of the year.

Governments elsewhere are also accelerating their move towards regulating generative AI tools, including China, Britain and several individual European countries, while others are seeking input on future governance.

Amid growing concern of generative AI, the head of German-based media giant Bertelsmann, Thomas Rabe has told the Financial Times that, on balance, it represented more of an opportunity for the creative industries.

This is provided that AI's potential and threats are understood and managed.

However, he acknowledged that the industry was in "uncharted territory" when it comes to protecting the intellectual property of authors and artists.

Rabe also pointed to the opportunities for authors and artists to augment their work by training AI tools using their own back catalogue, which could open up new paths in content creation.

Earlier this month, Spotify removed thousands of tracks created by an AI music streaming start-up and withheld associated royalties. It had come under pressure from music industry incumbents such as Universal Music, the FT reports.

Bertelsmann is already deploying AI across a number of its businesses. This includes Fremantle, which uses AI to dub TV shows like The X Factor into Arabic, while BMG and RTL Netherlands are looking to AI to help place adverts that fit with the content.

Market.us has forecast that the global market for generative AI in music will increase more than tenfold between 2022 and 2032.

Indian researchers have created an AI bot that helps make information on government programmes available in ten local languages across India. A mobile assistant available through Whatsapp, the bot could help overcome the language barrier in India, where government information is published in English although the language is only spoken by a tenth of the population.

Chinese search engine Baidu will launch its generative AI large-language model "very soon", CEO Robin Li announced on 26 May. Ernie 3.5 will power Baidu's ChatGPT-like app Ernie Bot and upgrade its search engine, Reuters reports.

Meanwhile, TikTok is conducting tests with users based in the Philippines on a "Tako" chatbot, which can help them discover "entertaining and inspiring" video content.

Microsoft has added live search results from Bing to ChatGPT, the viral chatbot created by its partner OpenAI. This means that ChatGPT can now draw on live Bing results, addressing the bot's previous limitation to information up to 2021.

It comes as OpenAI raised $175,250,000 for its start-up investment fund. Its original target was $100,000, underlining investors' ravenous appetite for everything AI. To date, 14 investors have contributed to the fund, and a second fund is in the works.

As AI is becoming more sophisticated, one big question is which jobs it will and won't be able to replace. Asked for its view, ChatGPT concluded it is unlikely to replace jobs requiring human skills such as judgement, creativity, physical dexterity and emotional intelligence. It is therefore not surprising that the latest Future of Jobs 2023 report expects the highest growth rates among agricultural equipment operators, drivers of heavy trucks and buses and vocational education teachers.

It's time to get serious about the AI ethics, standards and guardrails all of us must continue adopting and refining, writes Francesca Rossi, AI Ethics Global Leader, IBM. AI is an ever-evolving technology, so training must be part of a responsible approach to building both the AI systems in use today and those coming online tomorrow.

A recent survey finds that business leaders in North America significantly underappreciate AI risk. Many organizations are underprepared for AI, lacking proper oversight and expertise to manage the associated risks. This means AI can become a blind spot, preventing its ethical and effective deployment and overshadowing the transformative opportunities associated with its growth. At the World Economic Forum's Growth Summit 2023, experts explored how organizations can prepare more proactively for AI.