Companies prepare for disputes as AI use increases

Many companies expect a rise in legal disputes over the use and implementation of artificial intelligence (AI) and other technologies over the coming few years – even as they predict they’ll increase their spending in the area. In-House Perspective assesses what legal teams will face.

According to a survey of over 500 corporate counsel, risk managers, IT specialists and compliance professionals conducted by law firm CMS, half of respondents believe that the use of AI technologies will give rise to risks and disputes that cannot be foreseen now, and 56 per cent expect AI to be a leading source of increased disputes.

Respondents also said legal/compliance teams have a higher confidence in identifying current technology risks than executives do.

The survey, Technology Transformation: Managing Risks in a Changing Landscape, says the leading sources of disputes over the past three years have been compliance and regulatory issues (as stated by 65 per cent of respondents); performance/service level and outsourcing disputes (61 per cent); intellectual property (IP) and trade secrets (52 per cent); and software licensing disputes (51 per cent).

However, over the next three years, more than three-quarters of respondents expect compliance and regulatory disputes to decrease, while two-thirds see a decline in cyber breach disputes. Some 62 per cent expect IT performance or service level disputes to decrease.

Lee Gluyas, a partner in CMS’ dispute resolution team, says ‘the picture that emerges is of current risks being downgraded in their importance as organisations focus on prioritising how to manage emerging risks relating to new technologies’.

Respondents believe IP and trade secrets will become the biggest area for potential disputes in the coming three years, with issues arising from AI, smart contracts and cryptocurrencies following suit.

Consequently, dispute resolution should change to meet the risks of new technologies, say the majority of respondents, though only a third expect that disputes arising from AI technologies would follow the same principles as non-AI disputes.

Generally, arbitration was the most favoured, with 29 per cent of respondents ranking it in their top two choices, though 24 per cent still prefer litigation. Negotiation without a mediator is the least favoured process.

However, the popularity of litigation and arbitration varies across regions. In EMEA, 36 per cent favoured arbitration compared to just 14 per cent favouring litigation. In other regions, the two processes have broadly equal support.

Sajai Singh, Co-Chair of the IBA Technology Law Committee and co-chair of the Corporate Practice at J Sagar Associates, agrees that there are several circumstances under which possible legal disputes involving AI could arise. Key among these, he says, are commercial transactions involving AI, as well as relating to product liability.

‘As companies increasingly integrate AI into their products and systems, AI’s potential to cause injury and property damage is growing,’ says Singh.

‘AI’s ability to act autonomously raises novel legal issues,’ he explains. ‘An overarching question is how to assign fault when the product at issue incorporates AI. Since no law addresses explicitly injuries caused by AI, litigants and courts have begun to test the applicability of traditional legal theories involving injuries involving AI products. Product liability principles such as negligence, breach of warranty and strict liability would be applicable and tested.’

Singh adds that organisations using personal information for AI may face other legal challenges – for example, when attempting to comply with international data protection laws. Meanwhile, creating and using AI and related technology, such as computer hardware, can present unique issues regarding the protection of AI through IP, ownership of AI IP and AI IP infringement.

Competition rules could also be infringed if AI technology facilitates price-fixing agreements among competitors or creates anti-competitive agreements with other AI systems, says Singh.

Another area where risks can occur is the use of AI in the workplace. Here, its deployment, particularly around employee valuations and recruitment, increases the likelihood that discrimination and bias risks can occur in the screening, interviewing and hiring process.

Singh says it’s difficult to attribute particular fault to technologies and how they’re used because the decisions they take become more and more removed from any direct programming and are, in turn, based more on machine-learning principles. This makes it hard to attribute the question of fault and pinpoint the liability.

As a result, Singh believes dispute resolution will remain an important mechanism to settle complaints between companies. However, he adds that courts are largely unused to dealing with such cases.

Specific dispute resolution mechanisms could be more widely adopted to counter this issue.

An online dispute resolution (ODR) mechanism, which builds upon the foundation characteristics of alternate dispute resolution, is a cost-effective option as it relies on video conferencing and technology to transmit information. It’s also useful for resolving cross-border disputes and issues arising because of multiple jurisdictions – a common element where AI/emerging technologies are involved.

A flexible dispute resolution mechanism, along the lines of the new Digital Dispute Resolution Rules in the UK, could be introduced to resolve disputes involving novel technologies such as crypto assets, smart contracts and distributed ledger technology. The UK Rules provide a forward-thinking and innovative approach to digital dispute resolution that is fast, flexible and cost-effective, says Singh.

These regulations follow the government-backed UK Jurisdiction Taskforce’s recognition of crypto assets as property under English law and smart contracts as legal and enforceable.

Meanwhile, automatic dispute resolution is a unique concept introduced in the UK which allows an arbitrated outcome to be implemented directly by the arbitrator on a blockchain using a private key. For example, the arbitral tribunal may mandate that the losing party pay compensation automatically into the winning party’s digital wallet within a blockchain.

The tribunal may use any digital signature, cryptographic key, password, or other digital access to operate, edit, sign, or cancel any digital asset relevant to the dispute. It can also order any party to take such steps. Any disagreement settled by automated dispute resolution is binding on the parties. However, automated dispute resolution to digital assets is not mandated under the UK Rules. The parties can adjust the Rules’ application to their dispute and opt to enforce an award through conventional processes.

‘Modern dispute resolution strategies have shifted from litigation to arbitration. Because of their flexibility and consensual character, alternative dispute resolution processes are taking precedence over the arbitration,’ says Singh.