What AI can and cannot do

The risk of job replacement in public affairs

I had a friend in the public affairs space tell me this week that ChatGPT has transformed the way he works. He has become much more productive and has become an evangelist for AI in his workplace. This, three months after he first expressed ignorance on all matter of AI, and was skeptical that it had any role in his work.

He has the right mindset on this now. Long-time subscribers know that I firmly believe that AI can transform public affairs, and that those of us who learn how to use it effectively now, will reap career benefits later.

However, I’ve also found that many of us are uncertain about what role we’ll have with AI, and what it means for job security.

I’ve had younger colleagues express concern that AI could hamper their career growth. Part of growing in your career is taking on “grunt” work to understand how the sausage is made, learn your craft, get feedback, learn, and get better. What happens when senior people turn to AI instead of them? What learning opportunities are they losing out on?

Are there jobs being replaced?

How much longer until more senior jobs are replaced?

These are tough questions to answer. None of us really knows. But I have some working theories.

I’m guided by Kai-Fu Lee’s original thinking on AI’s job replacement potential. In his must-read book, AI Super-Powers, Lee offers this:

AI’s biases don’t fit the traditional one-dimensional metric of low-skill versus high-skill labor. Instead, AI creates a mixed bag of winners and losers, depending on the particular content of job tasks performed. While AI has far surpassed humans at narrow tasks that can be optimized based on data, it remains stubbornly unable to interact naturally with people or imitate the dexterity of our fingers and limbs. It also cannot engage in cross-domain thinking on creative tasks or ones requiring complex strategy, jobs whose inputs and outcomes aren’t easily quantified. What this means for job replacement can be expressed simply through two X-Y graphs, one for physical labor and one for cognitive labor.

The Y-axis moves from “asocial” at the bottom to “highly social” at the top. The X-axis moves from “optimization-based” on the left, to “creativity- or strategy-based” on the right. “Optimization-based” tasks involve maximizing quantifiable variables that can be captured in data (e.g. setting an optimal insurance rate).

The grid has four quadrants, with the following descriptors for cognitive labor:

• Human Veneer: computational work can already by done by machines, but the key social interactive element makes them difficult to automate at scale.

• Safe Zone: likely out of reach of automation for the foreseeable future.

• Danger Zone: high risk of replacement

• Slow Creep: work that relies on creativity or ability to adapt to unstructured environments. These remain substantial hurdles for AI, but only for now.

This grid gives us a heuristic for understanding what kinds of jobs are at risk, but what does this mean for public affairs practitioners?

Here’s my thinking on where the most common positions in public affairs fit in the grid. I’m not pretending there is a huge science to this. These are my best guesses, based on my understanding of these roles, and the automation potential in each one.

I provide a detailed summary and rationale for each position below, but first some broad observations:

Broad Observations

I believe large-scale analysis can be outsourced to large language models. It may take some technical know-how to get started, but whether this is done through an AI expert or by the savvy analysts on your team who see the writing on the wall, there is a way forward for jobs in the Danger Zone quadrant:

• The move for anyone in analyst positions is to learn to master large-scale analysis quickly and become the conduit between humans and LLMs.

Most jobs that we comfortably assume only a human can do, especially those that are low-level or mid-range in most corporate hierarchies, will be easily replaced by AI within the next decade, if not sooner. These are the jobs I’ve placed in the Slow Creep quadrant. Many of these jobs currently entail large levels of low-value work than we’d like to admit, and could easily be outsourced to AI.

• The move for anyone in these positions is to learn how to use AI agents to take on multiple roles and thus operate in an aggregate role. Rather than have a lobbyist, a grassroots campaigner, and a public policy consultant on your team, one person will take on all three roles because they know how to use three different agents with extreme precision.

There are jobs that are social in nature and rely on computational work to be effective. We will continue to be comforted by having a human being represent these roles, even if the bulk of the work is done by machines. I believe jobs that inherently involve human contact, will still require humans to make that contact. I’ve placed this jobs in the Human Veneer quadrant.

• The move for anyone on these positions is to lean into AI to make the analysis side of the job hyper-efficient, and using the gains in efficiency to become exceptional at human contact, something that will feel incredibly fresh in an automated world.

And then there are the jobs that require high levels of strategic thinking, people oversight, business leadership, all of which will no doubt be powered by AI, but ultimately will still need bright minds to oversee. Some jobs will just always require a human, and I think the most strategic roles will be safe for the foreseeable future. These are the jobs I’ve placed in the Safe Zone.

Let me know where you disagree in the comments section.

---

Analysis and Rationale Quadrant and Role

Danger Zone

1. Research Analyst: Interpreting data in context requires human insight, but data collection, processing, and statistical analysis can be optimized with AI. Risk of replacement: High.

2. Social Media Manager: Creative content creation and community engagement remain human-centric (e.g., crafting engaging posts), but scheduling, basic content generation, and analytics can be optimized with AI. Risk of replacement: High.

3. Policy Analyst: Interpreting nuanced policy implications requires human judgment, but data collection, statistical analysis, and drafting policy briefs can be optimized with AI. Risk of replacement: High.

4. Political Risk Analyst: Providing nuanced advice requires human judgment, but data collection, statistical analysis, and risk reports can be optimized with AI. Risk of replacement: High.

Slow Creep

1. Corporate Affairs Manager: Strategic elements remain human-centric (e.g high-level strategic decision-making and stakeholder negotiations), but many tasks can be optimized with AI (e.g monitoring industry trends, regulatory changes. Risk of replacement: medium.

2. Public Affairs Specialist: Creative communication and relationship-building remain human-centric (e.g., crafting tailored communication plans, engaging with diverse audiences), but routine tasks can be optimized with AI (e.g., monitoring media coverage, drafting standard press releases). Risk of replacement: Medium.

3. Internal Communications Manager: Facilitating employee engagement and fostering culture require human touch (e.g., leading difficult conversations), but distributing content and collecting feedback can be optimized with AI. Risk of replacement: Medium.

4. Public Relations Manager: Strategic planning and crisis management remain human-centric (e.g., developing PR campaigns, managing media relationships), but monitoring media and generating standard materials can be optimized with AI. Risk of replacement: Medium.

5. Brand Manager: Crafting brand narratives and creative content remains human-centric (e.g., developing brand strategies), but analyzing engagement metrics and distributing content can be optimized with AI. Risk of replacement: Medium.

6. Public Relations Manager: Crisis management and strategic campaign development remain human-centric (e.g., handling public image), but media monitoring and standard communications can be optimized with AI. Risk of replacement: Medium.

7. Lobbyist: Personal influence and negotiation with lawmakers remain human-centric (e.g., direct lobbying), but gathering policy data and preparing materials can be optimized with AI. Risk of replacement: Medium.

8. Crisis Communications Manager: Human judgment and empathy are crucial (e.g., making decisions during crises), but monitoring crisis indicators and preparing templates can be optimized with AI. Risk of replacement: Medium.

9. Cause Marketing Associate: Building partnerships and engaging stakeholders emotionally are human-centric (e.g., organizing events), but market research and tracking performance can be optimized with AI. Risk of replacement: Medium.

10. Public Relations Specialist: Managing media relationships and crafting PR strategies remain human-centric (e.g., organizing PR events), but monitoring media and distributing standard materials can be optimized with AI. Risk of replacement: Medium.

11. Corporate Communication Advisor: Providing strategic advice and acting as media spokesperson are human-centric (e.g., leading communication initiatives), but monitoring issues and writing standard materials can be optimized with AI. Risk of replacement: Medium.

12. Publicist: Crafting unique publicity campaigns and managing media relationships are human-centric (e.g., organizing promotional events), but distributing press materials and scheduling appearances can be optimized with AI. Risk of replacement: Medium.

13. Public Affairs Consultant: Strategic advice and navigating political landscapes remain human-centric (e.g., advising clients), but monitoring legislative developments and policy research can be optimized with AI. Risk of replacement: Medium.

14. Communications Consultant: Crafting persuasive messages and strategic planning remain human-centric (e.g., message development), but monitoring media trends and generating standard materials can be optimized with AI. Risk of replacement: Medium.

15. Regulatory Affairs Consultant: Interpreting complex regulations and advising strategy remain human-centric (e.g., negotiating with regulators), but monitoring changes and preparing compliance documents can be optimized with AI. Risk of replacement: Medium.

16. Issue Manager: Developing response strategies and engaging with stakeholders are human-centric (e.g., handling emerging issues), but monitoring media and analyzing trends can be optimized with AI. Risk of replacement: Medium.

17. Advocacy Consultant: Crafting compelling messages and persuading stakeholders are human-centric (e.g., leading advocacy efforts), but analyzing policy landscapes and managing databases can be optimized with AI. Risk of replacement: Low.

18. Public Policy Consultant: Strategic policy advice and relationship-building remain human-centric (e.g., engaging with policymakers), but conducting policy research and drafting documents can be optimized with AI. Risk of replacement: Medium.

19. Grassroots Campaigner: Motivating volunteers and personal engagement are human-centric (e.g., organizing community support), but managing databases and scheduling events can be optimized with AI. Risk of replacement: Low.

Human Veneer

1. PR and Events Manager: Personal networking and creative event planning are human-centric (e.g., building relationships during events), but logistical coordination and promotional material creation can be optimized with AI. Risk of replacement: Medium.

2. Media Relations Specialist: Building media relationships is human-centric (e.g., personal outreach to journalists), but monitoring media coverage and distributing press releases can be optimized with AI. Risk of replacement: Medium.

3. Government Relations Specialist: Personal interactions with officials and advocacy are human-centric (e.g., building relationships, negotiating), but tracking legislative developments and preparing reports can be optimized with AI. Risk of replacement: Medium.

4. Community Outreach Manager: Personal engagement and building community relationships are human-centric (e.g., hosting events), but managing contact databases and sending communications can be optimized with AI. Risk of replacement: Medium.

5. Fundraising Manager: Building relationships with donors and personal appeals are human-centric (e.g., securing major gifts), but donor data analysis and automating campaigns can be optimized with AI. Risk of replacement: Medium.

6. Investor Relations Officer: Building investor confidence and handling sensitive inquiries are human-centric (e.g., coordinating shareholder meetings), but preparing financial reports and analyzing data can be optimized with AI. Risk of replacement: Medium.

7. Government Relations Consultant: Personal lobbying efforts and building relationships with officials are human-centric (e.g., direct engagement with lawmakers), but tracking legislative activities and preparing reports can be optimized with AI. Risk of replacement: Medium.

8. Stakeholder Engagement Consultant: Facilitating dialogues and building trust are human-centric (e.g., engaging stakeholders), but mapping stakeholders and scheduling can be optimized with AI. Risk of replacement: Medium.

Safe Zone

1. Chief Communications Officer: Executive leadership and strategic decision-making remain human-centric (e.g., directing communication strategies), but data analysis and message standardization can be optimized with AI. Risk of replacement: Low.

2. Public Affairs Director: Strategic planning and stakeholder relationship management remain human-centric (e.g., developing outreach strategies), but monitoring external environments and reporting can be optimized with AI. Risk of replacement: Low.

3. Communications Director: Leadership and strategic oversight require human judgment (e.g., overseeing communication strategies), but media monitoring and content distribution can be optimized with AI. Risk of replacement: Low.