The success of a life sciences company relies on its research and development (R&D) and maintaining a robust pipeline of clinical products. Essential factors in building that pipeline start with capturing innovation from the scientists, ensuring that individuals consider their work valuable, and then reporting on their discoveries. Incentivizing scientists to think about and foster a culture of innovation is a complex task, as everyone is busy these days just completing their regular day to day projects.
Making this even more challenging is a question on the forefront of life science leaders' minds: Can artificial intelligence (AI) make our jobs easier, and if so, will we lose the human component essential to the innovation culture?
Here, we will explore how AI can speed up everyday processes without sacrificing creativity and camaraderie among internal team members. We also will explore how the U.S. government sees inventorship in the age of AI and share tips for team leaders and IP managers to foster innovation.
Are Pharmaceutical Companies Using AI To Its Full Potential?
While some pharmaceutical and biotechnology companies have built platform technology around AI and drug discovery, many companies are just starting to determine how and where AI can be beneficial. The good news is that companies embracing AI are still developing inventions and filing patent applications, underscoring that AI does not stifle innovation and may even help bring it along faster.
For example, programs for in silico modeling to determine whether a compound can bind, activate, or inhibit a target have been used for years. Thanks to AI, they are now hundreds of times faster and can simultaneously analyze thousands more conceptual compounds.
Before the computer starts whirring, teams typically brainstorm what ideal compounds might look like and what properties would be most beneficial. These grassroots development efforts can require more cross-department collaboration at the beginning of a project than may have historically been required and foster more fluid communication between different development teams. After the computer provides the results, they need to be analyzed. This, again, can facilitate more cross-department innovation, for example, around the optimal next steps in the development and drug approval process, e.g., biological efficacy, companion diagnostics, manufacturing processes, and discussions on possible other diseases in which the new compound can be effective.
Challenges with AI include the time and cost investment needed to implement any new system within a company and determine how to optimize its workflow. Team members may fear implementing an AI system will eliminate human jobs. However, explaining that AI is a time saver that is not designed to replace human ingenuity can temper reluctance to adopt the system.
Additionally, team members will be intimately involved in training the AI system to generate the correct findings and tweaking as needed to ensure the output is accurate and relevant. Team members' roles may shift due to AI, but they are still invaluable to maintaining the culture of innovation. Companies can ask where AI may assist in expediting internal processes and how to ensure these daily tasks are done easier, leaving more time for other, more creative-intensive tasks and brainstorming sessions across departments and development teams.
Questions On Inventorship: Patent Office's Safeguards On AI As An Inventor
The U.S. government has been considering AI's impact on many aspects of daily life. Its impact on innovation and inventorship has been tested, and guidance has been issued on how to handle AI in patent filings.
In a recent case before the United States Patent and Trademark Office (USPTO) and then the U.S. Court of Appeals at the Federal Circuit, a patent owner challenged the Patent Commissioner to allow an AI system to be listed as an inventor on a patent application. The USPTO ruled that an inventor must be a human, and an AI system cannot be an inventor (Thaler v. Vidal, Fed Cir 2022).
The government subsequently set out guidelines on how to treat inventions made with or assisted by AI and how a claim must be constructed to be valid if AI is involved.
The USPTO recognizes that an AI system — like other tools — may perform acts that, if performed by a human, could constitute inventorship under our laws. However, inventorship still requires the conception of the invention and a contribution to improve what may be already-known technology. If the subject matter is made solely by AI, then it cannot be properly claimed in a patent application unless there is human involvement in determining the subject matter of the claim.
For example, in drug discovery, an AI program (designed by a human, hint, hint!) can analyze chemical structures or other nucleotide or amino acid sequences through a program trained to identify a potential drug for a target. The computer presents the results to the development team. However, the team still must sift through the results, identify the best outcomes, determine their accuracy, and test them for efficacy, stability, etc. If an invention arises from that data, what is claimed will likely be the lead compounds that have activity, the discovery of which was made with human intervention and not just by AI.
Also, as hinted above, the AI has to be trained on what to look for and what parameters to use to identify compounds. Therefore, it can be argued that the person who taught the AI is an inventor because they knew what they were looking for, and the machine made their job easier.
The USPTO's position that AI cannot be an inventor is encouraging for companies that want to use AI but not inadvertently downplay the importance of their R&D teams in developing new therapeutics. Reminding scientists that AI is considered like any other research tool and that human involvement is necessary in analyzing the outcome of the computer research and the next steps in developing and using the patentable invention should bolster scientists' feelings that their contributions are paramount to what AI can generate.
Practical Advice
Maintaining the culture of innovation and scientists' morale within a biotech or pharma company can prove difficult when projects are ever-changing, and AI is emerging in a way that some could fear will replace their jobs. Reassuring your teams that AI is another faster research tool that still requires their scientific input and creative thinking is essential, as is recognizing those individuals responsible for training the AI to get the results and analyze the output. Fostering a culture of innovation with the emergence of AI still requires the internal reward and recognition of scientists' contributions to shepherding a new therapeutic through the different stages of development.
The USPTO's stance on AI as an inventor should also reassure scientists that AI cannot eliminate all creativity from drug development. According to the Executive Order on AI, inventorship is improper in any patent or patent application for which at least one natural person did not significantly contribute to the invention.
Companies should take extra care to ensure each named inventor in a patent application truly contributed significantly to a claimed invention. One way to do this is to add a question to the company invention disclosure form asking whether AI tools were used to create any part of the subject matter to be patented and how the AI was applied. Additionally, if AI was used, you should ask internally what were the next steps with the AI output, e.g., were further experiments performed to get to the ultimate invention, did one or more individuals train the AI to retrieve a specific finding such that the trainers are inventors, and who were those individuals?
Another question on AI in the patent process is whether and how one discloses to the patent office that AI tools were used to arrive at the subject matter. Is a reference in the specification sufficient? Is there an article or publication that can be submitted on an IDS relating to the AI system? The duty of disclosure is still in effect even if AI is in use, and this is another aspect to be aware of during the patenting process.
AI implemented by companies to speed up analytical processes can leave more time for collaboration and bring about new cross-department collaborations that may not have happened without advanced technology. Although AI may induce some growing pains within a company, AI should not be anxiety-inducing if scientists are reminded that their contributions are necessary for AI to produce the right results. They are the ones who truly analyze and vet the innovations that lead to drug discovery success.
Originally published by Drug Discovery Online.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.