The surge in AI, particularly with systems like ChatGPT, is facing a potential slowdown due to the impending depletion of publicly available text data, according to a study by Epoch AI. The shortage is projected to occur between 2026 and 2032, highlighting a critical challenge in maintaining the rapid advancement of AI.
AI’s growth has relied heavily on vast amounts of human-generated text data, but this finite resource is diminishing. Companies like OpenAI and Google are currently purchasing high-quality data sources, such as content from Reddit and news outlets, to sustain their AI training. However, the scarcity of fresh data might soon force them to consider using sensitive private data or less reliable synthetic data.
The Epoch AI study emphasises that scaling AI models, which requires immense computing power and large data sets, may become unfeasible as data sources dwindle. While new techniques have somewhat mitigated this issue, the fundamental need for high-quality human-generated data remains. Some experts suggest focusing on specialised AI models rather than larger ones to address this bottleneck.
In response to these challenges, AI developers are exploring alternative methods, including generating synthetic data. However, concerns about the quality and efficiency of such data persist, underlining the complexity of sustaining AI advancements in the face of limited natural resources.
Reporters Without Borders (RSF) has praised the Council of Europe’s (CoE) new Framework Convention on AI for its progress but criticised its reliance on private sector self-regulation. The Convention, which includes 46 European countries, aims to address the impact of AI on human rights, democracy, and the rule of law. While it acknowledges the threat of AI-fueled disinformation, RSF argues that it fails to provide the necessary mechanisms to achieve its goals.
The CoE Convention mandates strict regulatory measures for AI use in the public sector but allows member states to choose self-regulation for the private sector. RSF believes this distinction is a critical flaw, as the private sector, particularly social media companies and other digital service providers, have historically prioritised business interests over the public good. According to RSF, this approach will not effectively combat the disinformation challenges posed by AI.
RSF urges countries that adopt the Convention to implement robust national legislation to strictly regulate AI development and use. That would ensure that AI technologies are deployed ethically and responsibly, protecting the integrity of information and democratic processes. Vincent Berthier, Head of RSF’s Tech Desk, emphasised the need for legal requirements over self-regulation to ensure AI serves the public interest and upholds the right to reliable information.
RSF’s recommendations provide a framework for AI regulation that addresses the shortcomings of both the Council of Europe’s Framework Convention and the European Union’s AI Act, advocating for stringent measures to safeguard the integrity of information and democracy.
Meta has launched its first AI-driven ad targeting program for businesses on WhatsApp, aiming to generate revenue from the popular chat service. CEO Mark Zuckerberg announced the new tools at a conference in Brazil, marking a significant shift for WhatsApp, which has traditionally avoided targeted advertising.
The new AI tools will use behaviour data from Facebook and Instagram to target messages more effectively to users who are likely to engage, provided they use the same phone number across accounts. The new feature is crucial for businesses as it allows for optimised ad delivery, making their marketing efforts more cost-effective.
Meta is also testing a new AI chatbot for business inquiries on WhatsApp. Namely, the chatbot will handle common requests like finding catalogues or consulting business hours, pushing towards automated customer service solutions. Additionally, Meta is integrating Brazil’s popular digital payment method, PIX, into WhatsApp’s payment tool, enhancing its functionality in the country.
These developments come as part of Meta’s broader strategy to monetise WhatsApp, which, despite its massive user base, has yet to contribute significantly to Meta’s overall revenue. The new initiatives are seen as steps to leverage WhatsApp’s extensive reach and user engagement for greater financial returns.
According to European banking executives, the rise of AI is increasing banks’ reliance on major US tech firms, raising new risks for the financial industry. AI, already used in detecting fraud and money laundering, has gained significant attention following the launch of OpenAI’s ChatGPT in late 2022, with banks exploring more applications of generative AI.
At a fintech conference in Amsterdam, industry leaders expressed concerns about the heavy computational power needed for AI, which forces banks to depend on a few big tech providers. Bahadir Yilmaz, ING’s chief analytics officer, noted that this dependency on companies like Microsoft, Google, IBM, and Amazon poses one of the biggest risks, as it could lead to ‘vendor lock-in’ and limit banks’ flexibility. These facts also imply the strong impact AI could have on retail investor protection.
Britain has proposed regulations to manage financial firms’ reliance on external tech companies, reflecting concerns that issues with a single cloud provider could disrupt services across multiple financial institutions. Deutsche Bank’s technology strategy head, Joanne Hannaford, highlighted that accessing the necessary computational power for AI is feasible only through Big Tech.
The European Union’s securities watchdog recently emphasised that banks and investment firms must protect customers when using AI and maintain boardroom responsibility.
Top officials at the US Federal Election Commission (FEC) are divided over a proposal requiring political advertisements on broadcast radio and television to disclose if their content is generated by AI. FEC Vice Chair Ellen Weintraub backs the proposal, initiated by FCC Chairwoman Jessica Rosenworcel, which aims to enhance transparency in political ads, whereas FEC Chair Sean Cooksey opposes it.
The proposal, which does not ban AI-generated content, comes amid increasing concerns in Washington that such content could mislead voters in the upcoming 2024 elections. Rosenworcel emphasised the risk of ‘deepfakes’ and other altered media misleading the public and noted that the FCC has long-standing authority to mandate disclosures. Weintraub also highlighted the importance of transparency for public benefit and called for collaborative regulatory efforts between the FEC and FCC.
However, Cooksey warned that mandatory disclosures might conflict with existing laws and regulations, creating confusion in political campaigns. Republican FCC Commissioner Brendan Carr criticised the proposal, pointing out inconsistencies in regulation, as the FCC cannot oversee internet, social media, or streaming service ads. The debate gained traction following an incident in January where a fake AI-generated robocall impersonating US President Joe Biden aimed to influence New Hampshire’s Democratic primary, leading to charges against a Democratic consultant.
This week, key players in the chip industry, including Nvidia, Intel, AMD, Qualcomm, and Arm, gathered in Taiwan for the annual Computex conference, announcing an ‘AI PC revolution.’ They showcased AI-enabled personal computers with specialised chips for running AI applications directly on the device, promising a significant leap in user interaction with PCs.
Intel CEO Pat Gelsinger called this the most exciting development in 25 years since the arrival of WiFi, while Qualcomm’s Cristiano Amon likened it to the industry being reborn. Microsoft has driven this push by introducing AI PCs equipped with its Copilot assistant and choosing Qualcomm as its initial AI chip supplier. Despite this, Intel and AMD are also gearing up to launch their AI processors soon.
Why does it matter?
The conference was strategically timed to precede Apple’s annual Worldwide Developers Conference, hinting at the competitive landscape in AI advancements. As the PC market shows signs of recovery, analysts predict a rise in AI PC adoption, potentially transforming how PCs are used. However, there needs to be more skepticism about whether consumer demand will justify the higher costs associated with these advanced devices, as the Financial Times reports.
Salesforce has chosen London for its first AI centre, where experts, developers, and customers will collaborate on innovation and skill development. The US cloud software company, which is hosting its annual London World Tour event, announced last year a $4 billion investment in the UK over five years, focusing on AI innovation.
Zahra Bahrololoumi, CEO of Salesforce UK and Ireland, highlighted customer enthusiasm for AI’s benefits while noting caution about potential risks. She emphasised the importance of trust in AI adoption, citing Salesforce’s Einstein technology’s ‘Trust Layer’ that protects customer data.
Moreover, Salesforce’s dedication to responsible AI goes beyond data security. Bahrololoumi emphasises the company’s commitment to making AI a force for good. Their message to customers and partners is clear as they are deeply committed to collaborating closely to ensure that the transformative technology of AI brings about positive impacts.
In a recent interview, Intel CEO Pat Gelsinger presented how Intel intends to remain a relevant actor in the Chinese chip market while also scaling up production in the US. This was done at the occasion of the Computex tech conference in Taipei, where Intel released its new Xeon 6 processor, destined for data centres. Its release comes at a time when tech giants are challenging Nvidia’s chip dominance.
Gelsinger aims to build an Intel foundry in the US after the organisation was incentivised to increase facilities in the US with as much as $8.5 billion in grants and $11 billion in loans under the CHIPS and Science Act. In its release, the White House stated this is a step towards ‘protecting national security’ and increasing US share of global chip production to ‘20% […] by the end of the decade’.
“The capital is critical. We said that we have to have economic competitiveness if we build these factories in the US, and that’s what the CHIPS Act has done. It’s created a level playing field if I were building a factory in Asia versus US,” Gelsinger said.
Why does it matter?
At the same time, Gelsinger reiterated the importance of the Chinese market to his company. “China is a big market for Intel today, and one that we’re investing in to be a big market for Intel tomorrow as well,” he said. Intel has been competing to catch up its global market share since 2017, when South Korea’s Samsung overtook it as the largest chipmaker in terms of revenue. Since then, Taiwan Semiconductor Manufacturing Company reportedly overtook Samsung in 2023.
The US Justice Department and the Federal Trade Commission (FTC) have agreed to proceed with antitrust investigations into Microsoft, OpenAI, and Nvidia’s dominance in the AI industry. Under the agreement, the Justice Department will focus on Nvidia’s potential antitrust violations, while the FTC will examine Microsoft and OpenAI’s conduct. Microsoft has a significant stake in OpenAI, having invested $13 billion in its for-profit subsidiary.
The regulators’ deal, expected to be finalised soon, reflects increased scrutiny of the AI sector. The FTC is also investigating Microsoft’s $650 million deal with AI startup Inflection AI. This action follows a January order requiring several tech giants, including Microsoft and OpenAI, to provide information on AI investments and partnerships.
Why does it matter?
Last year, the FTC began investigating OpenAI for potential consumer protection law violations. US antitrust chief Jonathan Kanter recently expressed concerns about the AI industry’s reliance on vast data and computing power, which could reinforce the dominance of major firms. Microsoft, OpenAI, Nvidia, the Justice Department, and the FTC have not commented on the ongoing investigations.
Meta Platforms is facing 11 complaints over proposed changes to its privacy policy that could violate EU privacy regulations. The changes, set to take effect on 26 June, would allow Meta to use personal data, including posts and private images, to train its AI models without user consent. Advocacy group NOYB has urged privacy watchdogs to take immediate action against these changes, arguing that they breach the EU’s General Data Protection Regulation (GDPR).
Meta claims it has a legitimate interest in using users’ data to develop its AI models, which can be shared with third parties. However, NOYB founder Max Schrems contends that the European Court of Justice has previously ruled against Meta’s arguments for similar data use in advertising, suggesting that the company is ignoring these legal precedents. Schrems criticises Meta’s approach, stating that the company should obtain explicit user consent rather than complicating the opt-out process.
In response to the impending policy changes, NOYB has called on data protection authorities across multiple European countries, including Austria, Germany, and France, to initiate an urgent procedure to address the situation. If found in violation of GDPR, Meta could face strict fines.
