The ongoing debate among some of the most influential tech entrepreneurs regarding the inherent risks of artificial intelligence (AI) appears not to have dampened enthusiasm in boardrooms for acquisitions of AI companies this year. Indeed, CB Insights reports that 2017 is shaping up to be a historic period for M&A involving AI companies with 60 deals reported in the first half of the year, compared with 78 M&A deals reported in the AI sector during all of 2016. Many of these deals involve well-established corporations acquiring AI startups that have recently been spun-out of university labs. Some of the key issues commonly found in these deals include ownership of IP, special employee concerns and the management of deal announcement in a highly charged media environment.

Ownership and Rights to Exploit IP

Perhaps in part because of the abstract and novel ideas around which AI startups are formed, many founders of AI companies are professors or PhD students who are working at or for universities at the time the company is formed. This close nexus to universities gives rise to a raft of issues that should be reviewed when considering the acquisition of an AI startup. Who owns title to, and rights to commercialize and exploit, the intellectual property at the heart of the startup is fundamental to the value of the startup and should be clearly established at the outset of the acquisition diligence. Research universities have written policies governing the ownership of IP created by faculty and students, and the policy that applied at the time that the IP was created should be reviewed carefully. Most research universities also have Technology Transfer Offices (TTOs) that assist the university (including faculty and students) to protect and license IP developed by the university and generally to guide the commercialization of such IP by private enterprises. The TTO of the applicable university should be consulted early in the diligence process if there is any uncertainty as to who owns title to IP purportedly owned, or used by, the AI startup, or as to the scope of the AI startup's rights to the commercialization or proceeds derived from the use of such IP.

Confidential Communications

The collaborative environment fostered by and amongst universities can be both an asset and a risk to an AI startup. Communication of ideas between universities without appropriate documentation in place can undermine the value of the IP in those ideas, even if they have been protected by patents or other registrations, and lead to disputes as to ownership of IP. Acquirors should look for confidentiality agreements that specify ownership of IP in shared information, including any IP that is jointly developed or conceived in the exchange of information.

Ongoing Relationships with the University

Because of their deep ties to universities, founders of AI startups who are faculty or students of universities often are reluctant to end their academic careers in order to exclusively pursue the AI startup, even after its sale to an established corporate buyer. Buyers should be careful to explore in detail the extent of any ongoing academic engagements by employees of the startup and the applicable university's conflict of interest policy, and should establish boundaries around time commitments and the nature of what information can be used in lectures or publications where the content is closely related to the business, or anticipated business, of the AI startup. There is no one-size-fits-all solution to this issue, and acquirors should be cognizant of the value that the collaborative university environment may represent to the development of the AI startup, even after acquisition. At the same time, however, it is imperative that acquirors clearly understand, and are mindful of, rights that universities may have to IP developed by the acquired company's employees who continue to be affiliated with such universities.

Other Employee Issues

Acquirors should also be careful to protect their investment by requiring key personnel to sign non-competition agreements, particularly in acquisitions where talent, rather than technology, is the most valuable asset being acquired. However, in the context of an under-funded, pre-revenue startup—as most AI startups are—acquirors should be aware of the limitations on enforceability of non-competition covenants, especially in jurisdictions like California which prohibit by statute most non-compete clauses on public policy grounds. Most jurisdictions will enforce a noncompetition covenant that is undertaken in connection with the sale of a business by a person who receives substantial consideration from such sale. Accordingly, if a relatively small portion of the consideration in the deal is flowing through the capitalization table to the key employee against whom the non-compete would be enforced, acquirors consider reworking the structure of the transaction so that a meaningful amount of consideration is directed to the key employees through their capital holdings in the company.

Public Announcement

Finally, public company buyers should consider the impact that the announcement of a pending acquisition of an AI-based company may have on its stakeholders. There is currently significant public interest in AI and all parties to a pending acquisition should be cognizant of how its disclosure may feed into the broader narratives that are developing around AI. Parties should consider a wide array of interested constituents, including (institutional and retail) shareholders, customers, strategic partners, vendors, and current and future employees. In larger and higher-profile deals, a comprehensive communications plan should be developed so that key persons are prepared to respond to questions about the deal in ways that develop the most desirable narratives.