Since Open AI’s launch of ChatGPT (Chat Generative Pre-trained Transformer) in November 2022, there has been a great amount of discussion regarding the use of generative AI in the legal profession. While definitions of Generative AI have some variance, it generally refers to deep-learning models that can generate high-quality text, images, and other content based on the training data.
Generative AI is not brand-new. Generative AI was introduced in the 1960s, but it was not until 2014-2015, with the introduction of generative adversarial networks (GANs), which involve a type of machine learning algorithm, that expanded creative capability was realized. Notwithstanding, while generative AI has only recently become a tool more widely used in legal practice, for a significant number of years AI- based tools have been used by the legal profession in tools such as major legal research platforms.
Despite Chat GPT-4 successfully passing the Uniform Bar Exam and the Multistate Professional Responsibility Exam it has been acknowledged that general AI-based tools may not be suitable for many legal tasks; AI legal assistants that are engineered to produce answers from known and reliable legal data and provide linked citations which allow legal professionals to verify information have experienced more widespread adoption by law firms. Notwithstanding the substantial expense associated with these generative AI-based tools, powered by large language models, there has been concern expressed by some these tools could "replace" paraprofessionals and even lawyers. However, in the ABA Task Force on Law and Artificial Intelligence webinar entitled The Implications for Generative AI on Legal Education, Suffolk Law School Dean Andrew Perlman opined, citing the recent article he authored with the assistance of AI, that lawyers should use the technology to enhance their work and to efficiently provide legal services. Yet he cautioned that legal professionals should not relinquish their professional responsibility to AI by adopting AI produced content without reviewing every word.
Legal professionals have noted concern about developing technology competence regarding the use of generative AI tools. Courts have sanctioned lawyers who have misused generative AI tools which hallucinated and reported cases that did not exist. Courts have also considered requiring certification of human review when attorneys use generative AI tools.
The ABA Legal Technology Survey Report has polled lawyers for their opinions about AI-based legal technology tools for several years. Interestingly, the 2022 and 2023 23 Legal Technology Surveys define artificial intelligence as "intelligence displayed by machines such as when a machine mimics human cognitive functions like reasoning, learning, or natural language processing." The definition makes no reference or distinction between what is classified as generative AI.
The 2023 Survey questioned attorneys across the spectrum of law firm sizes, regarding whether they were currently using artificial intelligence-based technology tools, seriously considering purchasing artificial intelligence-based technology tools or not interested in purchasing artificial intelligence-based technology tools. Respondents also had the option of responding that they did not know or didn’t know enough about AI to answer the question. Given the significant nature of the legal news coverage of generative AI development, It might be somewhat surprising that over half of the respondents (58.8%) indicated that they did not know of the interest in AI based tools or didn’t know enough about AI to answer the question. However, the 2023 results are similar to the results of the 2022 Survey in which 57.8% of the respondents indicated such a lack of knowledge.
|
|
NUMBER OF ATTORNEYS |
|||||
|
Total |
Solo |
2-9 Attorneys |
10-49 Attorneys |
50-99 Attorneys |
100-499 Attorneys |
500 or more |
Currently using artificial intelligence-based technology tools |
10.9% |
10.4% |
9.2% |
11.3% |
3.6% |
10.6% |
21.4% |
Seriously considering purchasing artificial intelligence-based technology tools |
9.8% |
4.2% |
11.5% |
10.3% |
17.9% |
6.4% |
14.3% |
Not interested in purchasing artificial intelligence-based technology tools |
20.5% |
31.3% |
28.5% |
15.5% |
10.7% |
4.3 |
7.1% |
Don’t know |
27.5% |
8.3% |
13.1% |
38.2% |
39.3% |
63.8% |
42.9% |
I don’t yet know enough about AI to answer this question |
31.3% |
45.8% |
37.7% |
24.7% |
28.5% |
14.9% |
14.3% |
|
100% |
100% |
100% |
100% |
100% |
100% |
100% |
Count |
440 |
96 |
130 |
97 |
28 |
47 |
42 |
Overall, the survey results tell us that attorneys and firms are close to evenly divided between those that are already using or are seriously considering purchasing AI tools (20.7%) and those who are not interested in doing so (20.5%). Of course, there are differences in attitudes expressed based on practice/firm size. Of particular interest is the increase in adoption and consideration of AI-based tools by solo and small firms (less than 10 attorneys). In 2022, just over 15% of small firms and only 3.7% of solo attorneys were using or interested in AI-based tools. In comparison, in 2023, 20.7 % of small firms and 14.6% of solo attorneys indicated use or interest in such tools. Clio’s 2023 Legal Trends Report also noted that solo and smaller firms expressed interest in using AI more within the next year and had fewer concerns about AI-based technology, while larger firms indicated that copyright issues and other risk were a cause of concern.
It may be that interest in AI-based tools is increasing because it is perceived that such tools are becoming more mainstream. Both the 2022 and 2023 Technology Surveys asked: When do you think artificial intelligence will become mainstream in the legal profession? In 2022, just over six percent of respondents indicated that AI is already mainstream and approximately twenty percent responded that AI would become mainstream in the next three years. In 2023, approximately four percent responded that AI is already mainstream, but almost forty percent of respondents indicated that AI would become mainstream within the next three years.
Whether AI-based tools are already mainstream or are to become mainstream in the foreseeable future, is it important that attorneys receive training? When asked how important is it to you to receive training/education on emerging technology (e.g. blockchain, AI), twenty-five percent responded that training is very important and approximately forty-three percent rated it as somewhat important. Interestingly, when reviewing how firm size affected responses, in firms consisting of 50-99 attorneys just in excess of forty percent of respondents indicated that training in emerging technology, including AI, was very important.
Unfortunately, the survey did not include questions regarding what type of training would be of interest. When discussing the usefulness of training involving generative AI tools, some knowledgeable legal professionals have suggested that lawyers and law students could benefit from training regarding prompting design. However, some representing legal research systems using generative AI, such as vLex’s Vincent AI tool, have suggested that lawyers need to be able to ask good legal questions and the tools and platforms should be responsible for adapting the efficient prompts. However, this is not a universally held belief. The debate is somewhat reminiscent of the debate regarding whether lawyers should also be able to code.
Perhaps, the seeming increase in interest of solo and small firm attorneys mentioned above is related to the perceived benefits of AI-based tools. Solo and small firm attorneys are increasingly concerned in saving time and increasing efficiency. As reflected in the chart below, saving time and increasing efficiency was identified as the most important benefit of AI-based technology tools by approximately forty-four percent of the respondents to the 2023 Technology Survey. Similarly, but at a lower rate, just in excess of twenty-five percent of the respondents to the 2022 Technology Survey identified saving time and increasing efficiency as the most important benefit of AI-based tools.
What do you see as the most important benefit that artificial intelligence-based technology tools could provide to law firms?
|
|
NUMBER OF ATTORNEYS |
|||||
|
Total |
Solo |
2-9 Attorneys |
10-49 Attorneys |
50-99 Attorneys |
100-499 Attorneys |
500 or more |
Saving time/increasing efficiency |
44.3% |
35.8% |
41.9% |
47.4% |
60.6% |
46.8% |
50.0% |
Document management document review |
8.3% |
3.2% |
7.8% |
7.2% |
14.3% |
10.6% |
17.5% |
Reducing cost |
3.0% |
1.1% |
4.7% |
4.1% |
- |
4.3% |
- |
Predicting outcomes/reducing risk |
3.0% |
4.2% |
3.9% |
3.1% |
- |
2.1% |
- |
Quality assurance |
1.6% |
2.1% |
.8% |
2.1% |
3.6% |
2.1% |
- |
No benefits |
4.8% |
9.5% |
5.4% |
3.1% |
- |
4.3% |
- |
Other |
.7% |
1.1% |
.8% |
- |
3.6% |
- |
- |
Don’t know |
9.6% |
10.5% |
6.2% |
10.3% |
3.6% |
12.8% |
17.5% |
I don’t know enough about AI to answer the question |
24.7% |
32.5% |
28.5% |
22.7% |
14.3% |
17.0% |
15.0% |
|
100% |
100% |
100% |
100% |
100% |
100% |
100% |
|
|
|
|
|
|
|
|
Count |
436 |
95 |
129 |
97 |
28 |
47 |
40 |
It is curious that almost a quarter of respondents to the 2023 Technology Survey indicated that they did not know enough about AI to identify the most important benefit that AI-based tools could provide. This even though more than seventy-one percent of those responding acknowledged a requirement under the basic competency requirement of jurisdictional enactments of professional conduct rules to stay abreast of the benefits and risks of technology. At least there is a decrease since 2022 of those indicating insufficient knowledge of the benefits of AI-based technology. In 2022, almost thirty percent of the responses acknowledged a lack of depth of knowledge to answer the question.
In contrast, only approximately fifteen percent of respondents indicated that did not have sufficient knowledge about AI to identify major concerns about implementing and utilizing AI-based technology. Some of the major concerns expressed included accuracy of technology (57.7%), reliability of technology (48.1%), data privacy and security (46.5%) cost to implement (28.8%), time to learn (22.9%) and difficulty changing processes (17.8%). Interestingly, the percentage of those identifying concerns regarding accuracy and reliability of the technology increased significantly from 2022 to 2023. In 2022, approximately 36 percent of respondents indicated that accuracy of the technology was a major concern, while almost thirty-four percent of the responses disclosed major concerns with the reliability of the technology. In 2022, cost to implement was identified by more than thirty-four percent of the respondents while data privacy and security concerns were significantly less than in 2023 with only approximately 27 percent of expressing concern about privacy and security matters.
It is encouraging to see attorneys taking advantage of available resources to learn about the risks and benefits of technological developments, including Artificial Intelligence. It is particularly heartening that the 2023 Technology Survey noted that CLE seminars and webinars are the most cited source of such information with Legal News rated as the second source. Other sources consulted included legal technology consultants, other law firms, publications, product vendors, podcasts, and trade shows (such as the ABA TECHSHOW). While some are troubled by vendors providing such information, some of the most informative presentations and resources I have attended and consulted regarding generative AI include vendor resources.
Generative AI tools were not used in the original preparation of this report. However, such tools were used in review. Select modifications recommended by these tools were adopted.
