HomeMy WebLinkAbout04.a Discussion on How to Incorporate the Use of Artificial Intelligence (AI) at the DistrictPage 1 of 7
Item 4.a.
F__1_448�411C_S0
December 5, 2023
TO: ADMINISTRATION COMMITTEE
FROM: CHARLES MALLORY, INFORMATION TECHNOLOGY MANAGER
REVIEWED BY: PHILIP LEIBER, DEPUTYGENERAL MANAGER -ADMINISTRATION
ROGER S. BAILEY, GENERAL MANAGER
SUBJECT: DISCUSSION ON HOW TO INCORPORATE THE USE OF ARTIFICIAL
INTELLIGENCE (AI)ATTHE DISTRICT
Artificial Intelligence (AI) has the potential to significantly improve Central San operations and help meet
the needs of customers in new ways, ranging from predictive analysis and maintenance to streamlining
communications and bolstering the support of cybersecurity. Al can also be used for cost savings,
enhanced service delivery, improved safety and security, fraud detection and prevention, data analytics,
and tracking regulatory compliance.
The use of Al within Central San should consider privacy and security, compatibility with legacy systems,
and evolving workloads. Al can streamline routine tasks and processes, free up staff to focus on higher -
value activities and deliver better services and outcomes to citizens.
Two main types of Al include (1) Generative Al, which is being used to generate novel content, including
text, images, videos, code and music, and (2) Predictive Al which is being used to predict future events.
The use of Al enhanced applications in local government and special districts are broad and growing.
Some examples of Al applications in government include:
• Hyper -personalized services: Al can provide personalized services to citizens, such as chatbots that
can answer questions and provide information.
• Office Efficiency- Generative Al tools (such as the increasingly popular Chat GPT) can be used to
develop initial drafts of documents such as policies and procedures.
• Proactive citizen engagement: AI -powered chatbots can engage citizens and provide them with
information about Central San services.
• Al can predict system failures, maintenance schedules, and system outages, which can help
prioritize repairs and cleanups.
• AI can monitor security cameras and identify potential security threats.
• Al can predict natural disasters and help prepare for them.
• Al can also be used to promote sustainability by optimizing energy usage.
December 5, 2023 Special ADMIN Committee Meeting Agenda Packet - Page 62 of 70
Page 2 of 7
It is important to note that Al adoption and maturity levels vary depending on the government agencies,
their existing infrastructure's reliance on legacy systems, and workforce fluency in these tools.
Some of the potential risks and concerns about Al include:
• Concerns aboutjob displacement: Al could potentially replace human workers in specific tasks,
leading to concerns about job displacement. However, if effectively managed with this concern in
mind, Al tools can instead lead to greater efficiency and allow workers to do their jobs better.
• Ethical concerns: Al can be biased and raise ethical, privacy, or discriminatory concerns.
• Security risks: Al can be vulnerable to hacking and other security risks.
• Lack of human -like creativity and empathy: Al lacks humans' creativity and empathy.
It is because of the aforementioned risks that Central San should develop an Al policy to govern its use to
protect the employees, customers, and the organization. The goal is to obtain the benefits of Al, while
avoiding or mitigating the potential for the risks noted above. The policy should allow and encourage the
continued use of Al in administration and operations while providing guidelines to avoid misuse and the
sharing of sensitive information, and to continue to center human judgment in core decisions related to
business operations. The policy should include definitions and usage guidance, and should be reviewed
on a regular basis to update and to stay current with technological and ethical developments. Other
government agencies have begun to develop Al policies, and Central San staff will use these as a source
for developing a policy specific to the needs of Central San. Staff will return to the Administration
Committee/Board with a draft policy for further discussion and potential approval.
Strategic Plan Tie -In
GOAL SEVEN: Innovation and Agility
Strategy 1 - Leverage data analytics to become a more efficient utility, Strategy 2 - Implement organization -wide
optimization
ATTACHMENTS:
1. Presentation
December 5, 2023 Special ADMIN Committee Meeting Agenda Packet - Page 63 of 70
Page 3 of 7
1
Let's start with the word of the year:
Artificial Intelligence (AI)
Machine Learning (ML):
Machine learning is a subset of Al that focuses on developing algorithms that enable systems to learn from and make predictions or decisions based on existing data. A subset of Al that
teaches the computer to learn from the data inspired by humans.
Accounting data
Facial recognition
Generative Artificial Intelligence (Al):
Al, machine learning algorithms are designed to learn and improve from experience without being explicitly programmed. They use statistical techniques to identify patterns, extract
insights, and make informed predictions.
A subset of machine learning focuses on creating models capable of generating new content that resemble existing data
Text, image, music, and video generation
• Computer Programming or Coding
Business simulation
2
December 5, 2023 Special ADM IN Committee Meeting Agenda Packet -Page 64 of 46 DRAFT 1
Page 4 of 7
Suggested Best Practices When
Using GPT-Based Al
Adhere to Central San's Acceptable Use Policy for PH protection
DO NOT share confidential information with GPT Al
Always fact -check using trusted sources 11
I
Monitor for copyrighted material usage
Treat generative Al as a tool to enhance your work, not replace it!
3
4
December 5, 2023 Special ADM IN Committee Meeting Agenda Packet -Page 63 of 46 DRAFT 2
Page 5 of 7
Al and Communications
Programs used:
Adobe Beta testing program: Firefly Al, Photoshop Beta
Adobe Al Tools: Photoshop, Lightroom, Illustrator
Grammarly Go
Benefits:
Develop ideas and concepts
Enhance image quality and streamline photo editing
Editing text content for publications
5
Al and Communications
Before After
All generated sign
_ �.- base and frame.
HOUSEHOLD
,:
Operating Schedule
Mon Sat. 700, -to 2M—, -L
kri f—at-go fa:
7:OOa.m. to 2:00p m.
�' MpO'NDAY— SATURDAY
eanbal.en.arg/hhw
a
�I
1
December 5, 2023 Special ADM IN Committee Meeting Agenda Packet -Page 66 of 48 DRAFT 3
Page 6 of 7
7
Aquasight Apollo Process Advisor Pilot Test
Operations Optimization Division is leading a pilot
test for Aquasight's Apollo process advisor tool on
the two processes shown here.
The tool will use real-time data analysis (including
artificial intelligence and machine learning) to:
Evaluate process performance;
Suggest changes to improve energy efficiency and
chemical usage; and
Identify opportunities for preventative
maintenance to try to avoid emergency fixes.
Pilot testing results are expected in June 2024.
I
Treatment Plant
Influent Pumping
NEW
Solids Dewatering
Centrifuges Polymer Dose
December 5, 2023 Special ADM IN Committee Meeting Agenda Packet -Page 65 of 46 DRAFT 4
Page 7 of 7
December 5, 2023 Special ADM IN Committee Meeting Agenda Packet -Page 66 of 46 DRAFT