Artificial intelligence (AI) is changing how people work, communicate, learn, and organize. Corporations and state institutions use it to process information, automate tasks, monitor workers, and strengthen their power. The Communist Party USA cannot stand outside these changes. As interest in socialism grows, the Party must turn that interest into stronger organization, political education, and working-class action.

The National Committee has argued that building a mass Communist Party requires mastering modern methods of communication and organization while continuing patient, face-to-face work among working people. For an organization with limited funds, few paid staff, and work carried mainly by volunteers, digital tools, including carefully governed use of AI, can be a force multiplier. This article argues for a coordinated approach across appropriate areas of Party work, without giving up human judgment, collective leadership, security, or direct ties to working-class struggle.

AI is only one part of a wider digital system. Many Party needs are better served by ordinary software, secure databases, shared documents, websites, and automated workflows. AI becomes useful when it helps people work with information inside those systems. Marx offered a useful starting point when he wrote that “general social knowledge has become a direct force of production,” describing how accumulated human knowledge becomes embodied in productive machinery. AI is one present form of that process, but the Party should still choose the right tool for each problem rather than force AI into every task.

Two wrong approaches on the Left

One mistaken response is technological worship. It treats AI as an independent force that can solve poverty, planning, education, or political organization through technical progress alone.  A weaker version appears when organizations adopt AI because it is new, trust fluent answers too readily, or confuse faster output with better political work. AI can process information, but it cannot create class consciousness, settle political lines, build trust, or accept responsibility. Lenin’s warning about scientific management remains relevant: “the domination of capital converts all these devices into instruments for the further exploitation of the workers.” Technology separated from working-class politics can strengthen surveillance, bureaucracy, exploitation, and war.

The opposite mistake is technological abstention. It begins with real dangers: corporations use AI and algorithmic systems to monitor and manage workers, transform or eliminate jobs, gather personal data, and weaken worker control. Concerns also include biased outputs, copyright disputes, the energy demands of data centers, and military adoption. Yet refusal will not prevent capital or the right from using these systems. Lenin confronted a similar contradiction in the Taylor system: he denounced its brutality under capitalism while insisting that a socialist government must “adopt all that is valuable in the achievements of science and technology,” adapting those achievements to different social ends. A Communist approach must therefore oppose harmful uses, fight for working-class control, and carefully use appropriate tools to strengthen Party organization and mass work.

The development of AI is not taking place outside the class struggle. A handful of powerful corporations control the infrastructure, the fruits of technical expertise, and online platforms. The development of AI under Capitalism raises serious questions about workers’ rights, surveillance, intellectual property, energy consumption, and accountability. Communities facing the construction of enormous data centers have legitimate concerns about electricity, water, and the use of public resources. Communists should stand with workers and communities demanding environmental protections, accountability, and a say so over technological development. Understanding these contradictions, however, does not require us to abandon every useful application of the technology. The question is how we can use this technology to build the Party and wage the class struggle.

An organizational problem

Working-class organizations face a deep resource imbalance. Wealthy political parties, foundations, think tanks, and campaigns employ paid staff, consultants, research teams, communications departments, databases, and large technology budgets. The CPUSA carries out many of the same basic functions with far fewer resources, largely through volunteers who also have jobs, families, union duties, and community responsibilities. Clubs and districts must welcome members, collect dues, prepare meetings, organize education, respond to events, maintain communications, preserve records, and support mass struggles. Too often, this work rests on a few reliable comrades using separate spreadsheets, scattered email threads, and personal memory. Tasks are delayed, useful skills remain unknown, and lessons disappear when experienced members become overworked or step away. The problem is not commitment; necessary work exceeds the organized time, staff, and technical capacity available.

Coordinated software, automated workflows, and carefully governed AI can help the Party use that capacity more effectively. Ordinary software can track dues, prepare reminders, organize records, and map broad geographic patterns. AI can draft communications, summarize approved material, support research and translation, and explain patterns for human review. An opt-in skills directory could connect needs with comrades willing to help in areas such as education, technology, bookkeeping, and labor outreach. None of these systems should decide membership status, assign comrades without discussion, or rank people’s political value. Their purpose is to reduce repetitive labor and give members more time for political education, cadre development, mass organizing, and direct contact with working people.

Major political players already treat digital infrastructure as basic organizational capacity. The Democratic National Committee provides campaigns and state parties with shared voter databases, analytics, data warehouses, application interfaces, documentation, and contact tools. The Republican National Committee reported that it shared voter-scoring technology with candidates and supplied an app combining organizational, fundraising, and political functions. These official accounts show what the major parties have built; they do not by themselves prove that every system is effective. They do show that powerful political organizations develop common infrastructure across campaigns instead of leaving technical work to scattered individual efforts.

What makes a party intelligent?

An intelligent party is not simply one that uses advanced technology. It learns from its work and from working-class struggle, connects local experience to Marxist analysis, preserves lessons, shares knowledge, develops cadre, and corrects mistakes without panic or factionalism. Its intelligence belongs neither to one leader nor to one group of experts; it grows through the organized exchange of theory, experience, discussion, decision, action, criticism, and correction. The CPUSA Constitution describes collectivity as the Party’s basic style of work and gives clubs a role in shaping broader policy while applying national strategy to local struggle.

Lenin’s description of a revolutionary newspaper helps explain the organizational role technology can play. A newspaper, he wrote, was not only a propagandist and agitator but “also a collective organiser,” a structure that helped comrades communicate, divide work, and see the common results of organized labor. Today’s digital systems can serve a related purpose by making approved reports, decisions, educational material, and organizing lessons easier to find and share. AI may assist with summaries and comparisons, but it must not replace discussion or decision-making. A Party becomes intelligent when technology helps its members think, learn, and act together more effectively.

AI as a force multiplier

The most important Communist work still requires people to do the heavy lifting. Organizing workers, welcoming members, building trust, leading education, resolving disagreements, and developing cadre depend on judgment and direct contact. These tasks are slow because they involve real people, contradictions, and responsibilities. Yet every hour spent finding documents, reformatting material, or recreating earlier work is an hour unavailable for organizing. AI, working alongside sound software and workflows, can help with first drafts, research, translation, summaries, and other supporting tasks. The goal is not more pressure on volunteers, but more time for the work only people can do.

Evidence from political organizing shows both the value and limits of digital tools. A nationwide field experiment found that text reminders increased voting by 3 percentage points. A 61-million-person experiment found that online mobilization affected real-world voting, with much of the effect moving through close social ties. But an $8.9 million campaign-wide experiment found no average turnout increase from broad digital advertising and concluded that even large campaigns were likely to have modest effects. The lesson is not that technology works or fails in the abstract. Its value depends on the task, the design, and whether it strengthens organized human relationships.

Consider a Party club whose officers work full-time jobs and have only a few hours each week for administrative responsibilities. Preparing a meeting might involve reviewing publicly available materials, developing an agenda, locating educational resources, and preparing announcements. This work is straightforward but can take a lot of time. With the right digital tools, club cadres could assemble a first draft of an agenda, summarize official educational materials, translate public announcements, or create emails in a fraction of the time that it would take without AI. Those tasked with the work would still review the material and make corrections before passing it out.

Suppose these tools saved a club’s cadre just two hours each month. Across fifty clubs it would represent 100 hours returned to the organization every month. Those hours could be spent contacting members, preparing educational discussions, supporting labor struggles, or building relationships in the community. Small changes in routine work, multiplied across an organization, can make for a significant increase in organizational capacity and time.

What AI cannot do

Political and organizational disagreements arise at every level. Clubs may differ over local priorities, districts may balance competing needs, and national bodies must decide under changing conditions and incomplete information. Differences in experience, tactics, communication, or follow-through can sharpen these contradictions. They cannot be solved merely by producing more information faster. The Party must distinguish principle from secondary difference, judge what conditions have changed, and determine what course best serves the working class. That requires patience, trust, accountability, honest speech, and a willingness to correct mistakes.

AI cannot resolve these contradictions because it has no class commitment, political responsibility, lived experience, or real relationships with the people involved. It may summarize documents or organize competing arguments, but it cannot determine political line, judge good faith, repair trust, or answer for a collective decision. These tasks require Marxism-Leninism to analyze concrete conditions; democratic centralism so members can discuss fully, decide collectively, and unite in action; and good-faith, comradely discussion in which criticism strengthens people and the organization. AI may support the edges of that process, but it cannot replace the political work through which a Communist Party develops judgment, unity, and trust.

There is an important difference between AI-assisted political work and giving AI authority over organizational and political matters. A committee may use AI for research and summarize different points of view. The AI’s output must be checked against reliable sources and evaluated by humans. AI-generated conclusions can help start conversations but cannot substitute for the political judgment of a collective. No AI system should independently create Party policy, develop political lines, or participate as a decision-maker in a Party collective. The authority to develop, debate, adopt, and implement political positions belongs exclusively to the Party’s members and its established bodies. 

Some principles for responsible digital and AI use

If the CPUSA expands its use of AI and other digital tools, it should do so through collective discussion and political oversight. The Party should identify the work to be improved, decide whether AI is actually needed, keep sensitive information outside inappropriate systems, name the people responsible for review, and assign authority to the proper Party bodies. This risk-based approach is consistent with the NIST framework for generative AI. Open-source software should be preferred where practical and reliable, while volunteers, consultants, and service providers may each have a role. No essential system should depend on one company, platform, or comrade.

The choice is not simply between a large commercial provider and a system built entirely by the Party. Routine records and workflows may be best handled by ordinary open-source software. Commercial services may be appropriate for low-risk work using public information. More sensitive work may require stronger privacy, limited access, or tools controlled more directly by the Party; some tasks should not use AI at all. Membership lists, disciplinary matters, confidential correspondence, credentials, legal strategy, and sensitive personal information should never be entered into public consumer AI systems. The Party should match each task with the safest useful tool.

AI models and providers differ in privacy, cost, ownership, control, and support. Technically knowledgeable comrades should explain those options to the appropriate collective bodies in clear political and practical terms. Those bodies, not technical experts alone, must weigh security, reliability, cost, access, maintenance, and political purpose. 

Before adopting an AI tool, a Party organization should answer several basic questions. What information will the system receive? Who can access it? Where will that information be stored? Can the provider retain or reuse it? What happens if the system produces inaccurate information or is compromised? Who is responsible for checking its work?

Sensitive membership information, confidential correspondence, disciplinary records, and internal deliberations should not be submitted to public AI services. AI systems should not receive unrestricted access to Party membership databases or other sensitive organizational records. Where a proposed use involves confidential information, it should require a separate security assessment and authorization by the appropriate Party body. Even locally operated systems require safeguards against unauthorized access, information leaks, and unreliable outputs. 

The starting point should be the least sensitive task, using the simplest reliable tool.

From individual skill to collective capacity

The CPUSA already includes technology workers and other comrades skilled in software development, cybersecurity, data analysis, web design, communications, technical support, and systems administration. That knowledge is a Party resource, but it should not remain scattered or concentrated among a few people. When only one comrade understands a system or controls an account, individual skill becomes organizational weakness. Technical comrades can evaluate tools, improve workflows, train members, document systems, and develop common political, ethical, and security standards. Clubs and districts need room to adapt tools locally, but essential work should be documented and shared. The goal is not a priesthood of technicians. It is to place specialized knowledge under collective oversight and turn individual skill into lasting Party capacity.

AI will not build the Communist Party USA, determine its political line, resolve contradictions among comrades, or organize the working class. Those tasks require Marxist-Leninist analysis, democratic centralism, collective leadership, and participation in struggle. Yet the Party cannot ignore digital systems already reshaping work, communication, knowledge, and political organization. Under clear limits and political purpose, software can strengthen routine operations while AI assists selected research, writing, education, and communication tasks.

One useful approach is to test small, low-risk uses and let experience guide broader decisions. A Party body might begin with a routine task involving no sensitive information, then ask whether the tool saved time, produced accurate results, protected Party information, and preserved the intended political message. Practical tests can help collective bodies develop standards based on evidence rather than fear or enthusiasm alone.

The Party doesn’t have to start with the latest AI models and tools. We can begin by learning which work consumes the most time from clubs, districts, commissions, and individual comrades, where existing tools fall short, and which new tools can help. Some solutions will involve AI, others only software and clearer procedures.

These questions should be explored collectively. Comrades with technical experience can help evaluate possibilities, while organizers, educators, and club cadres can identify the needs that technology should serve. Small experiments, conducted with safeguards, can provide lessons for the wider Party. Successful approaches can be documented, shared, and improved. Unsuccessful ones can be abandoned without great expense.

This article is a starting point for discussion throughout the Party. Its proposals must be tested against the needs and experiences of our clubs, districts, and national bodies. As William Z. Foster reminded us, “Organization does not happen; it is made to happen.” The same is true of our collective capacity. No machine can build the Party for us, develop our political judgment, or form the class solidarity that only comes through struggle. But we can learn to use the tools of our time to strengthen the work we do together. The question before us is not whether artificial intelligence will transform society, but whether we can put its useful capabilities to work without surrendering the principles that make us Communists.

The opinions of the author do not necessarily reflect the positions of the CPUSA.

Images: AI and the CPUSA. Fred Barr / CPUSA; Computing system inside the Scientific Data and Computing Center (SDCC). U.S. Dept. of Energy on Flickr. Public Domain; Lisa A. presenting workshop at 32nd CPUSA National Convention. Taylor Dorrell / People’s World; Wim Klerkx. Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 Netherlands.