The power of artificial intelligence (AI) is fundamentally changing the way in which corporations operate. Healthcare providers are using it to process medical scans and x-rays, online retailers are using it to analyze customers and generate buying recommendations, and car manufacturers are using it to develop autonomous vehicles. Few industries will be untouched by AI, and many will be transformed by it.
There has been a flurry of deal activity as companies look to acquire AI talent and technology that can be embedded into their products and services.
According to CB Insights, the number of acquisitions of AI start-ups climbed 44% in 2017 to 110 deals, up from 80 in 2016. Activity has continued apace in 2018, with notable deals including Amazon’s purchase of threat detection start-up Sqrrl and Oracle’s acquisition of Zenedge, a provider of cloud-based, artificial intelligence-driven cybersecurity solutions. Both purchases had the underlying motivation of tackling complex cybersecurity threats facing businesses.
The most active M&A acquirors are large tech firms—those possessing the financial muscle to snap up smaller start-ups with advanced AI capabilities, and at a high price. For such firms, AI is the perfect tool to analyze the huge amounts of data they collect in new, faster, cheaper and more accurate ways.
Apple, for example, acquired Emotient, an AI developer that is working on a tool using facial recognition technology to measure customer reaction to advertisements. And music-streaming business Spotify bought machine-learning startup Niland and its AI platform that processes user playlists to generate content recommendations.
In what is still a relatively nascent sector, with many companies holding great potential but still pre-revenue, it is unsurprising that venture capital firms have been active investors, too. CB Insights estimates that around 42% of the AI companies acquired since 2013 have had venture capital backing. According to Stanford University’s AI Index venture, investment into companies developing AI systems is up six-fold from 2000 levels.
Crossing sector lines
Although technology companies and venture capital firms dominate, the buyer universe for AI companies is widening. Companies operating in sectors other than tech are recognizing AI’s potential and are investing accordingly. In February, Roche Pharmaceuticals announced its US$1.9 billion purchase of Flatiron Health, a health-tech group that makes use of AI technology to gather and analyze patient data that then feeds into cancer research workstreams.
As a rapidly growing area, the autonomous car industry is ripe for M&A activity. This year has been characterized by considerable investment as firms look to beef up their technological expertise in the area. Ford Motor, for example, announced the creation of a self-driving car unit in July, and is planning to invest US$4 billion by 2023.
This follows a flurry of M&A activity in 2017, which saw, among other deals, Delphi acquire autonomous driving start-up nuTonomy in October for US$400 million. Dealmakers are on the lookout for the next big transaction.
AI is also transforming the way in which financial services companies operate. Japan’s Mizuho Financial Group expects that AI could replace up 19,000 staff by 2027, UBS has installed Amazon’s digital assistant Alexa on customer service, and HSBC has installed AI to detect money laundering and terrorist funding.
AI businesses have developed niche, domain-specific applications able to solve problems facing a particular industry. As a result, the growth of “non-core” AI is attracting buyers aside from traditional tech firms. Results International estimates that only a third of AI activity in 2017 involved “core” AI (autonomous systems, driverless cars and computer vision), with the other deals involving a much broader range of acquirors and applications.
In healthcare, for example, research has shown that machines are better at detecting skin cancer than dermatologists are. Companies in the healthcare space are taking note, as seen in a partnership between Nuance, a radiology image exchange network used by 70% of US radiologists, with NVIDIA, a deep learning platform, to create an AI marketplace for diagnostic imaging.
Transforming the deal process
For deal professionals, AI is not just an exciting source of new transaction flow. It can now be applied to every part of the deal process, from tracking and sourcing deals through to due diligence, execution and post-deal integration.
Starting with tracking and deal sourcing, AI can automate certain tasks and smooth workflows. Thomson Reuters has developed an AI product that can predict which companies will be M&A targets by analyzing text and patent content. Meanwhile, a study by EY observes that private equity firms are applying statistical algorithms to analyze company, target and third-party data to identify targets, speed up deal execution and inform M&A decision making.
The due diligence process has proven particularly suitable for the application of AI. Rather than hiring huge teams of people to sift through all a target company’s employment, supplier and customer contracts, AI platforms such as Kira, RAVN, eBrevia and Luminance search thousands of uploaded contracts across hundreds of data points. This enables them to present any issues to legal advisers and due diligence providers in a fraction of the time with at least the same level of accuracy. Due-diligence start-up Neotas uses AI to run background checks on management teams by searching the entire internet, including public records and social media, for any issues or red flags.
Systems such as Contract Express, Neota Logic and High Q, meanwhile, can be used to draft contracts, saving time and reducing the incidence of human error. When it comes to post-deal integration, AI can process huge volumes of emails and documents and help both parties involved in the deal to sort and integrate their respective data sets.
Potential legal issues
Deal professionals increasingly need to consider the unique legal issues arising from acquisitions of AI businesses.
For example, AI systems typically need to mine large sets of data in order train their underlying algorithms for developing products or services. Such data sets may be subject to copyright and other use restrictions. As an AI system “learns” from an input data set, the system could inadvertently produce resulting technology that could be deemed an unauthorized derivative work, and buyers may have trouble determining through due diligence whether the AI developer has the rights to use the third-party intellectual property (if any) contained in the data set.
If data sets include personally identifiable information whose use is restricted by privacy laws, then products and services derived from that data may also be impacted. And although products liability law is generally well-established, it is yet unclear how liability should be allocated for accidents, malfunctions and errors that result from decisions made by AI systems.
Breaking new ground
Just as AI is sparking a wave of M&A across all industries, it is also overhauling the way those very deals are done. The combination of companies needing to keep pace with how their businesses can implement AI, and the recognition among AI developers that they need to focus their technology on serving specific industry needs, have resulted in increased M&A activity across nearly all sectors. Deal professionals will do well to note this trend and familiarize themselves with the unique legal issues that this burgeoning field may present.