AI essentials: MIT reality check, the sustainability paradox, US deregulation and EU Code of Practice

MIT report delivers sobering reality check

A new report published by the Massachusetts Institute of Technology (MIT) has found that most organisations’ GenAI initiatives across a wide range of functions have failed to deliver the rapid revenue growth initially anticipated. Just 5% of AI pilots have achieved meaningful revenue acceleration, with the majority showing little measurable impact. According to the findings, the problem isn’t model performance but flawed enterprise integration.

Companies that succeed typically focus on solving a clear pain point and partner strategically, while many stumble by pursuing costly proprietary builds or over-investing in sales and marketing tools with limited ROI. The real value is emerging in less visible areas such as automation of back-office processes, reduced outsourcing and streamlined workflows.

These findings echo our own research, which shows that the true impact of GenAI still remains unclear. Most brand marketers are still at early stages of implementation however, with success hard to measure given the difficult in identifying the right KPIs.

The report, which was published on August 19, had an immediate impact on Wall Street, with US tech stocks suffering their sharpest sell-off in months as investors responded to MIT’s data and warnings of an AI ‘bubble’.

WFA investigates the AI sustainability ‘paradox’

As AI adoption in marketing accelerates, the sustainability impact can be overlooked. Brands are facing a potential paradox: while AI offers opportunities for more sustainable innovation (e.g. more efficient product development), it can also simultaneously introduce potential new environmental costs (e.g. energy-intensive content creation).

To understand how brands are approaching this dichotomy, we have launched a short survey exploring views on the environmental impact of AI use in marketing, whether they are factoring this into their adoption of AI and how they believe AI can support their company’s overall sustainability goals. You can take part in this survey here.

The findings will be revealed and discussed at an upcoming remote WFA AI x Sustainable Marketing Community meeting on November 5, open to all WFA members.

US signals shift towards deregulation

The White House has unveiled a national AI framework aimed at boosting US competitiveness and accelerating AI adoption. The plan focuses on expanding data centre capacity, removing regulations that ‘hinder AI development and deployment’, promoting US AI exports and awarding government contracts to developers whose systems are ‘objective and free from ideological bias’.

While the framework is not binding, it signals a federal push to prioritise speed and innovation over new regulatory controls, leaving most rule-setting at a state level.

While this approach may speed AI adoption within marketing, it also means ongoing legal uncertainty, particularly around issues relating to transparency in AI models, IP and copyright issues, data protection and privacy. Individual brands will have to define AI practices in a way that safeguards their ability to communicate responsibly and effectively.

Generative and Agentic AI are coming for media

New WFA research reveals that generative and agentic AI is set to have a major impact on brands’ media practices. The research, based on input from senior media leads across 37 multinational brands, finds that every respondent plans to integrate the technology into their media operations.

For the moment, the report found that media strategy, optimisation and buying are the priority areas and 90% believe it will enable faster reporting and greater efficiencies.

However, the enthusiasm comes with significant caution. Sixty-five per cent of respondents believe that agentic AI will increase risks, with top concerns including data protection and privacy, transparency and accountability and over-reliability on external tools.

WFA’s Media Forum will be unpacking both the opportunities and the risks in the months ahead. To participate in the conversation, contact WFA’s Global Lead, Media Services, Tom Ashby.

Meta refuses to sign EU AI Code of Practice

Meta has announced it will not be signing the EU’s voluntary AI Code of Practice, a framework published in July to help ‘general purpose AI providers’ such as OpenAI, Google and others prepare for compliance with the EU’s AI Act. Under the code, signatories commit to keep up-to-date documentation on their AI tools and services, avoid training AI systems on pirated material and respect requests from content owners to exclude their works from training datasets.

Meta’s Chief Global Affairs Officer Joel Kaplan claims the code introduces ‘legal uncertainties’ and measures that ‘go far beyond the scope of the AI Act’. OpenAI, Google, Amazon and several other US providers have already signed on.

Knowing that potential partner is a signatory to the Code of Practice could offer a degree of reassurance to individual marketers when they evaluate AI tools and providers.

WFA brings AI training to public affairs professionals

The potential of GenAI isn’t limited to marketing, it also offers powerful new ways to shape advocacy, influence policy and engage stakeholders. To help public affairs leaders make sense of and capitalise on these opportunities, WFA will be hosting an exclusive 90-minute training session in collaboration with specialist training company Influence Builders.

The training, taking place remotely on October 7 and open to members of the WFA Policy Forum, will explore real-world use cases, tools and strategies that can help policy professionals apply AI in core aspects of their public affairs work, from messaging to framing and persuasion, all while upholding ethics and transparency.

Please send across any tips, developments and interesting insights to Gabrielle Robitaille.